The Indian trade regime

Policy Research WORKING PAPERS Woa 0 %C, Country Operations Country Department The World Bank October 1992 WPS 989 The Indian Trade Regime M. Ataman Aksoy Reforming export policies alone - without reforming India's import and tax systems - will produce only marginal improve- ments. The whole system needs rationalizing. Policy Rcewrch WorkingPapersdisseminate the findimgsof work in progrews and encourage the exchange of ideas among Bank staff and a1 others interested in development issues.Thesepapers, ditfibuted by thelReseatrchAdvisory Staft,cury the names ofthe authors.dlect only theirviews,and shouldbe used and cited acoordingly.Thefindings, interptutions,and conclusions au theauthoreown.They should na be attributed to the World Bank, its Board of Diractors, its management, or any of its member countries.  Policy Research Contry Operations WPS 989 This paper - a product of the South Asia Country Department III -is based on a larger study of India's trade regime undertaken by the Department. Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Rose Matenda, room J 11-217, extension 35055 October 1992, 82 pages). Despite attempts to liberalize India's import accurately for their tax burdens. The system that trade regime, the structure of import licensing is exists is far too complex. still restrictive and complex and for most prod- * The absolute level of tariffs on inputs must ucts, the licensing systems probably offers no be reduced to administer the duty-free import more protection than tariffs do. For most prod- schemes efficiently. High tariffs encourage ucts, trade restrictions are probably redundant as leakage of duty-free imports into the domestic protection. market and abuse of high drawback rates (incen- Reforming export policies alone - without tives). reforming India's import and tax systems - will * Tariffs and taxes on capital goods must be produce only marginal improvements. Problems reduced to reduce the costs of investment. Tariffs in the export administration can be resolved only in India - especially on key intermediate by making changes in four areas: products (metals and chemicals) and capital * The import licensing system must be goods - are high and getting higher fast. The rationalized to eliminate import restrictions on high cost of basic inputs increases the cost of inputs and components. The import regime production, leads to uneconomic import-substitu- inflicts heavy administrative costs on the Indian tion which causes pressure for more protection, economy. Imports of raw materials and other and requires an elaborate, cumbersome system to inputs essential for production are delayed, compensate exporters. High tariffs and excise leaving downstream producers idle when domes- taxes on capital goods damage Indian competi- tic supplies are interrupted (which happens tiveness, adding 1(0 to 15 percent to the cost of often). The export regime is still not rationalized production and severely handicapping exporters. for smaller producers, indirect exporters, and Thb excessive tariffs do not fulfill their firms that rely on domestic suppliers. primary purpose of providing protection and * Tariffs and excise taxes must be consoli- incentives; they are aimed mainly at generating dated around two to three slabs and the quantita- revenues. Public revenues should be generated tive restrictions in intermediate and capital goods through more efficient instruments, especially must be eliminated so firms can be compensated taxes. | TPedPolicy Research Working Paper Series disseminates ioefindingsof work under way in theBank. Anobjectiveof theseries is to get these findings out quickly, even if presentations are less than fully polished. The findings, interpretations, and conclusions in these papers do not necessarily represent of ficial Bank policy. Produced by the Policy Rese-arch Dissemination Center  THE INDIAN TRADE REGIME BY M. ATAMAN AKSOY* * Principal Economist, India Department, Southern Africa Department, World Bank. This Paper is based on a larger study of India's trade regime undertaken by the India Department. This study does not reflect the major reforms that have been implemented since July 1991. The mapping of items in the import policy document to the Harmonized System (HS) was undertaken by Madhumita Gupta. Umnuay Sae-Hau drafted the section on tariffs and prepared the database used for the tables. Helena Tang prepared the quantification of export incentives. The staff of ICICI, especially the late Dr. Sanpath S. Iyengar, supplied the data on export profitability. The officials in the Department of Revenue and CCI&E contributed immensely to the establishment of the database. The input-output table used as --the base was constructed by Prof. M. R. Saluja of the Indian Statistical Institute. The study also drew heavily from the earlier work of Garry Pursell and Donald Keesing. Numerous helpful comments were received from Michael Gould, David Greene, Francois Ettori and Ashok Khanna.  TABLE OF CONTENTS Paae No. A.THEIMPORTLICENSINGREGIME .......................... . 1 The Existing Import Control System . .*..... . .3...... Quantitative Assessment of the Import Regime . . . . . . . . . . . 10 B. THE STRUCTURE OF TARIFFS . . . . . . . . . . . ...... . . . 17 The Existing Tariff Structure . . . . . . . . . . . . . . . . . . . 20 Evolution of Tariff and Excise Tax Collection Rates . . . . . . . . 23 Tariffs and Quantitative Restrictions . . . . . . . . . . . . . . . 27 Excise Taxes and Excisable Output . . . . . . . . . . . . . . . . . 29 C. EXPORT POLICIES AND ADMINISTRATION . .c...... eseesse........ 31 Export Controls . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Export Incentives . . . . . . .... . .... . .... . .... . .... . ... . . 32 Quantification of Export Incentives .... . . . . . . . . . . . . 38 Impact of Trade Regime on Export Profitability . . . . . . . 43 D. CONCLUSIONS . . . . . . . . . . . . a * * * * * * * * . * . * X e e 49 APPENDIX A: MATCHING IMPORT, HS AND INDUSTRIAL CLASSIFICATIONS . . . 55 APPENDIX B: DISAGGREGATED DATA BASE FOR TARIFFS . . . . . . . . . . . 71  1. Despite its importance in the post-independence growth of the Indian economy, the study of the trade reginie has received relatively little attention. After the pioneering work of Bhagwati and Srinivasan (1975) and Panchamukhi (1978) which basically covered the experience of the 1960s, the study by Wolf (1982) focused only on exports. The experience of the 1970s and early 1980s was studied in an unpublished Ph.D. dissertation by Rao (1985). The description of the trade regime in the late 1980s is contained, again in unpublished monographs, by Pursell (1988) and Keesing (1987, 1988). Even in these studies, the complexity of the trade regime and the unavailability of up-to-date statistics have made it very difficult to quantify the extent of QRs, detailed tariff structures and the magnitude of export incentives. 2. This study tries to quantify the structure of the import licensing regime, analyze the detailed structure of tariff and excise taxes and their evolution and up date the nature and magnitude of export incentives. It is based on a large study carried out by the India department that mapped individual items in different import licensing categories to actual imports, obtained disaggregated information on tariffs, analyzed firm level data on export competitiveness, and organized this data around an input-output table built for 1987/88. This database allows us to quantify, on a consistent basis, the scope of QRs, tariffs and export subsidies. The impact of the trade regime on the historical behavior of exports, imports and industrial output is summarized in Aksoy and Tang (1991) and will not be discussed here. Similarly, the effect of the trade regime on effective protection and industrial efficiency is discussed in Aksoy and Ettori (1991). The descrip- tions of import and export policies draw heavily from Pursell (1988) and Keesing (1987, 1988). Section A covers the import regime, Section B, the tariff structure and Section C describes the export regime which compensates for the import and tariff regimes. A. THE IMPORT LICENSING REGIME 3. India has had a very restrictive trade regime since the late 1950s. The instruments that have been used to regulate import demand have been an exten- sive import licensing system and high levels of tariffs. The import control  -2- mechanisms in India were first introduced as e result of the foreign exchange crisis of the Second Plan (1956-61). From 196u to 1977 these controls were increasingly tightened and made more complex. By 1977, when foreign exchange reserves were considered adequate, the Government began to relax import controls and quantitative restrictions.l/ This import liberalization phase has lasted until 1990. 4. Some of the more important import policy changes since 1978 include the expansion of OGL (Open General License) lists, shifting of goods from more to less restrictive lists, swifter and less restrictive administrative Judgments and some reduction in the scope of canalization. In particular, there has been a relaxation of restriction on capital and intermediate goods imports, though primarily of commodities not competing with domestic production. Imports of most consumer goods continued to be banned. Most of these changes were made to allow particular domestic industries (including a number of export industries) to modernize. Capital and intermediate goods imports which were almost constant in real terms between 1966 to 1977 have risen substantially in recent years to more than 6X p.a. (Aksoy and Tang, 1991). 5. The complexity of the import regime makes it very difficult to quantify the impact and significance of QRs. First the product descriptions in different licensing categories vary in coverage from very specific to general and thus do not give any indication of the extent of items covered. Second, imports are not recorded by licensing categories. Third, classification of items in the import policy is not organized according to the Harmonized System (HS) classification used to report imports. Therefore, most of the previous analysis on the import regime have been descriptive and the changes had to be analyzed in terms of numbers of items shifted from different categories without measuring their impact on the magnitude of imports or products covered. I/ These changes were Incorporated in the trade policy statement of 1978/79. The basic change which applied to raw materials, manufactured inputs, consumer and spares, involved a switch from 'positive" lists of allowable imports to a system of "negative* lists of restricted imports. The complex policy regarding import of capital goods and technology was also liberalized but on a more ad hoc basis. There was no change in the policy which bans most consumable goods imports.  -3- 6. Attempts have been made to estimate the share of imports subject to different licensing categories. Pursell (1988) has made rough estimates of imports under OGL for the 1980s and quotes estimates made by the Indian Government for 1980/81. These estimates are, however, very broad and based on a series of key assumptions that are difficult to verify. 7. This section tries to estimate the structure of import licensing system more precisely for 1987/88, by individually mapping the items in the Import- Export Policy document to the corresponding HS codes at the six digit level. First, a brief description of the import regime, drawn heavily from Pursell (1988), is presented and then the quantitative results of the licensing system are analyzed. The Existing Import Control System 8. India has a very complex system of import licensing that was developed piecemeal over the last three decades. The basic document of import controls is a three volume Import and Export Policy Document issued annually until 1985, and now issued every three years. This document classifies items by licensing regime and describes the procedures for import. Numerous changes are made in this document from time to time by public notices, which make it very difficult to ascertain the scope and extent of the QRs. For example, in the first eight months of 1990, 102 import control and 60 export control public notices were issued. 9. Non-tariff barriers operating through the import licensing system have long been the principal means of regulating imports and protecting domestic industries. These controls include: (i) the import licensing system; (ii) *actual userw policy (which forbids imports by intermediaries such as wholesalers); (111) canalization (monopoly import by a public sector firm); (iv) phased manufacturing programs (that mandate progressive import substi- tution); (v) the industrial licensing system and (vi) government purchase preferences for domestic producers.  10. Import Licensing System2/. The basic structure of the import licensing categories can be summarized as follows: (a) A Banned List that contains a few commodities mainly for health or religious reasons; (b) A Restricted List of commodities that require an import license, and are usually not importable except under special circumstances; (c) A Limited Permissible List of commodities that aiso require a license less restrictive than the Restricted List; (d) An Open General License (OGL), category of items not requiring an import license; and (e) Canalized Products (mostly raw materials) whose import is restricted to public sector agencies or public sector manufacturing enterprises. To some extent, items can be in different lists (i.e., OGL, limited permissible) and also be canalized. Then the canalization agencies import these items under different import regimes. 11. In addition to these major categories, some of these items can be imported by exporters through e,,port related licenses. The most common of these is the tradable Replenishment (REP) licenses, which allow exporters to import items on Canalized and Limited Permissible lists for own use or resale. 12. The import licensing system further subdivides imports into three broad categories by type of item: (i) consumer goods; (ii) capital goods; and (111) intermediate raw materials, components, spare parts and supplies. 13. Finished consumer qoods have long been in the Restricted list, and are practically banned with the following exceptions: (a) a few consumer goods in short supply imported by canalizing agencies, including edible oils, cereals, kerosene and certain drugs, etc.; (b) personal effects in passengers' baggage and gifts up to modest value limits; (c) medical and dental equipment and supplies used by doctors, dentists, and hospitals; (d) a few consumer goods are on OGL; i.e., pulses; books and Journals in educational, scientific and technical fields; educational films; dried dates; wheelchairs; homeopathic 2/ This is a highly stylized and simplified description of the import regime which consists of 17 appendices, each with various separate lists.  -5- medicines; some computer sottware; photographic film: and (added in the new Policy) four spices--nutmeg, cloves, cinnamon and mace--which were formerly canalized; (e) a few industrial inputs on one of the limited penmissible lists have potential non-industrial uses (paints, paper stationery, mirrors, locks, rubber hoses, and ethyl alcohol). 14. Capital goods imports are organized on a positive list principle and are divided into a *restricted" category and an "Open General License" (OGL) category. The import of capital goods that are not specifically on the OGL list are treated as restricted and require an import license. In addition, there is also a specific list of restricted capital goods which presumably indicate the items that an import license should not be granted. 15. OGL import of capital goods is subject to a number of conditions, of which the four most important are: (a) the importing firm must be the "actual user" of the equipment, which for five years cannot be resold without the permission of the licensing authorities; (b) the resulting change in produc- tive capacity must be compatible with the capacity approved by the industrial licensing authorities: (c) most items on the OGL machinery and equipment list are grouped according to the industries which predominantly uses them, and firms not belonging to these industries would need to obtain special permis- sion to import them; (d) special restrictions apply to imports of second-hand machinery and equipment, even if it would otherwise be on OGL. 16. Since 1976 there has been a steady increase in the number of capital goods on the OGL lists: From 79 in 1976 to 1170 in April 1988. This osten- sibly was done to allow domestic industries to modernize. However, most of the machinery put on the OGL list was not produced domestically so relaxation of the list might not imply competition with domestic producers. An exception is the machine tools where items domestically produced have also been placed in OGL. There also has been reduced stringency in the licensing of imports of the majority of machinery and equipment items which remain subject to discre- tionary controls. A global tendering facility is available for industries or projects in 14 important categories regardless of the availability of indig- enous capital goods, with selection of suppliers subject to scrutiny by the Empowered Committee. Also under "project imports" and technology development  -6- funds, imports of machinery have been allowed much more easily in the last few years. 17. One of the changes in the 1988-91 Import-Export Policy is to allow recognized Export and Trading Houses to import capital goods through the use of export related licenses (Additional Licenses). Evidently this rpeision was caused by concern over the costs and difficulties to small firms of importing directly. 18. Intermediate goods are divided into: (a) banned; (b) restricted; (c) limited permissible; (d) canalized; and (e) OGL categories. In general, intermediate goods which are not in one of the first three categories, nor on the separate canalized lists, can be imported on OGL, i.e. without a license. In practice, however, the system has not functioned according to this apparent "negative listo principle, mainly because of ambiguities in the lists of controlled items. Therefore, OGL status for intermediate goods have been limited to items included in the published OGL lists, plus informal internal OGL lists accumulated by the licensing authorities.3/ In any case, the negative list principle for intermediate goods was officially revoked in late 1990 due to balance of payments difficulties. It was not clear to which category these so called Oresidual" items were moved. However, import licenses are now officially required for intermediate goods not specifically included in the OGL lists. 19. As with capital goods, all intermediate goods imported under any of the four categories above fall under the actual user" requirement. The exceptions are provisions for imports of specified raw materials by the government canalizing agencies, a small list of raw materials and components 3/ Importing an unlisted product without a decision on whether it could be imported on OGL is not possible because the banks would not release the foreign exchange if there was any doubt. Therefore, the import of unlisted items could be more difficult than importing unrestricted or limited pe.miissible items, contrary to the Intentions of high level policy makers. The move to a negative list prlnciple was regarded as one of the most important changes recommended by the 1977/78 com1ittee, but the principle was never implemented, and instead case-by-case decisions were made on individual applications and informal internal lists were built up of those items which could be imported on OGL.  on OGL which can be imported for stock or resale by private importers, and raw materials which can be imported by virtue of special provisions applying to exporters (especially REP licenses) and to authorized rerchants in the export business ("export houses" and *trading houses") through Additional Licenses. 20. There has been a steady increase in the number of intermediate goods on the OGL list, and a reduction in the various banned and restricted lists since 1977. Some loosening of raw material controls has occurred in f ch year beginning with the 1977/78 policy. However, the main thrust of these policy changes has been to ease the supply situation of inputs used by domestic industries but not produced domestically. Most of the items put in the OGL category were either not made in India or not likely to be made in the coming year. In fact, each year some items are removed from the OGL to the various restricted lists on the ground that domestic products are available. 21. The main innovation in the 1988-91 and 1990-93 Policies, is that import replenishment (REP) licenses are now given for practically all exports (exclu- ding gems) and can be used to import any item in limited permissible and canalized lists. These licenses are transferable to anyone. Thus, traders and domestic manufacturers will be able to get on all the items on these lists by buying these licenses. 22. Non-OGL Imports. All non-OGL imports, except for those by exporters, are subject to a case-by-case decision. In each case, the "sponsoring agency" of each firm must certify to the Chief Controller of Imports & Exports (CCI&E) that the import is "essential," and an "indigenous angle clearance" must be obtained, usually from the DGTD (Directorate General of Technical Develop- ment), which certifies that a product of satisfactory specifications and quality cannot be supplied in a reasonable time by an Indian firm. For capital goods, approval must be given by the concerned Capital Goods Com- mittee. The cost of the domestic alternative is generally not considered in the decision. Other considerations include the foreign exchange availability, and the capacity ,pproved by industrial licensing authorities. In sum, import licenses are issued in a non-price, and administratively ad-hoc manner.  -8- 23. Canalized Imports. Canalizing agencies are another means by which the Government controls imports. There are 12 designated such agencies listed in the Import-Export policy. There was not always a clear distinction between canalized Imports and licensed imports, since some import licenses were issued to canalizing agencies to import apparently canalized products. During the late 1970s and early 1980s this overlapping occurred on a large scale. Import canalization was adopted three decades ago in India. Between 1969 and 1973, there was a steep growth in the proportion of canalized imports to total imports, an outcome of the Government's decision to be more interventionist with regard to foreign trade. The share of canalized imports increased from 37.28% of the total in 1968-69 to 65.81% in 1973-74. From 1980, the share of canalized products in total imports has declined substantially, from about 67% in 1980/81 to about 27% in 1988. The fall in shares was less a reflection of decanalization than it was the result of: (i) an increase in domestic POL production, a decline in POL imports and a decline in world prices of crude oil and petroleum products; (ii) the disappearance of grain and cotton imports; and, (Mii) large declines during the 1980s in the international prices of some of the other principal canalized imports. Nevertheless, there was decanalization of 21 items in April 1985 and a further 26 items in April 1988, although as a percentage of total canalized imports, these figures are rather small. 24. Actual User Policy. This policy excludes non-government intermediaries from lmporting, thereby disallowing imports for resale. The implications of this policy for different categories of imports are as follows: (a) consumer goods - Actual User Policy would prevent most of these imports even if the ban were lifted; (b) intermediate goods - special import licenses may be issued to allow canalizing agencies and certain private export and trading houses to import raw materials and components in bulk for resale to manufacturing fimrs. Also, private intermediaries can legally bypass this poli.1_ to some extent by acting as agents combining the replenishment licenses for exporters; (c) cap- ital goods - the policy greatly restricts imports of capital equipment and intermediate goods by small and medium firms, and even by large firms when the quantities required are small, owing to the high transaction costs of importing on a small scale.  -9- 25. The only items that can really be imported freely are items imported through REP licenses.4/ Anybody who buys a REP license can import the prod- ucts that are in canalized and limited permissible lists. Thus the most important distinction in this categorization is between the commodities in the restricted list and other categories. Items in other categories can be imported by REP licenses, while items in the restricted list cannot be imported by REP licenses. Although the proportion of imports through REP licenses are small (about 10 of exports), the potential import of these items allow the gaps between demand and supply of these commodities to be met by imports. For the commodities in the restricted list, adjustment of imports automatically to meet the excess demand in the domestic market is not possible. 26. Phased Manufacturing Programs (PMPs). These constitute a new devel- opment which runs counter to the general trend of the easing of restrictions on imports of intermediate inputs. The PMPs accompany investment licenses and involve agreement by the concerned firm to progressively replace imported materials, parts and components with those produced in-house or by other Indian firms. To ensure implementation of the agreements, the import of all such parts and components requires prior clearance (List Attestation) by the sponsoring authority for the industry. These agreements and procedures there- fore amount to a separate set of quantitative import controls which apply to many intermediate products, including those on OGL lists, which in theory are importable without the restrictions. 27. Industrial Regulatory System. This is in effect another discretionary, non-tariff barrier that protects the machinery and engineering industries against comp;ting imports. The system involves the clearance by the Capital Goods Committee (CGC) of applications for industrial licenses for new or expanded capacity. The CGC scrutinizes the foreign exchange content of the investment and may reject applications involving what it regards as excessive foreign exchange outlay, or may require local sourcing of particular machinery and equipment items. The latter is applicable even if the items are on the 4/ The only exception is the few items that are on OGL-stock and sale list. This list has about 75 items and imports in this category are less than 1X of total imports.  - 10 - OGL lists. Technology import policies also protect the c3iftal goods indus- tries. Under these policies, the allocation of foreign exchange to pay the royalties and license fees are subject to a case-by-case administrative review. The requirement of technology licenses for foreign firms, where such licenses are often linked to capital goods imports, is another form of non- tariff barrier to importing capital goods and technology. Finally, the reservation of particular products for exclusive production by small scale firms Is also an indirect import barrier. 28. There has been liberalization of the controls on domestic industry since 1985. Some of the policy changes pertaining to this are include: (i) in- creases in the asset limit below which firms do not need an industrial license: (ii) greater flexibility for increasing capacity without obtaining a license (through provisions such as ucapacity endorsement*); (iii) greater flexibility for diversifying production (through provisions for "broad banding" of industrial licenses). In addition, a number of product groups were delicensed, and in June 1988, a major reform carried this process further by delicensing all industries except for a negative list of 27 industries. Delicensing has freed up import restrictions in many ways: (i) only import policies, not industrial licensing policies, can in principle affect firms' decisions on whether to import capital equipment or buy it locally; (ii) with delicensing, there are no longer independent limits derived from industrial licensing which constrain the quantities which can be imported. 29. Government Purchase Preferences. Government purchases give preference to domestic firms. A price preference equivalent to 25X of the CIF price of imports plus duties and port charges is given to domestic suppliers by the Directorate General of Supplies & Disposals, which buys supplies for the central government and substantial amounts on behalf of state governments and state enterprises. Since Indian tariffs are high, this is equivalent to a. substantial margin on CIF prices. Quantitative Assessment of the Import Realme 30. Given the complexity of the import regime, it is very difficult to quantify the impact and significance of QRs in India. First, the descriptions in different licensing lists vary in coverage from very specific (Broach  - 11 - sharpening machine for grinding broaches up to 250 mm diameter and 2000 mm length and equipped with hydraulically operated cross slide for flat broach grinder) to very general (all consumer goods howsoever described, of indus- trial, agricultural or animal origin, not appearing individually in Appen- dices 3, part A and 5 or specifically listed for import under (pen General License). Thus, the number of entries does not give any indication of the extent of items covered. Second, the classification of items in the import policy is not organized according to the harmonized system (HS) classi- fication used to report imports and tariffs. So items, as specified in import policy, can not be directly compared to actual imports. And third, customs does not record imports by licensing category, so information on the magnitude of imports under different licensing categories cannot be directly observed. 31. For the purposes of this study, individual items in the 1988-1991 Import-Export Policy document were mapped into the corresponding HS codes at the six digit level. The methodology of the mapping is given in Appendix A. Given the variance in levels of disaggregation ir, items listed in the policy document, however, the mapping to HS codes are not exact. Furthermore, some of the HS codes are not mentioned at all in the import policy. Instead of making assumptions about items not listed in the policy document, they are treated as residuals that can not be classified. This partial mapping covers 82% of imports by value and 70% of items by number of six digit HS codes. 32. There are three methods of estimating the share of items in different licensing categories. First is the distribution of HS codes (more than 4,500 items) by licensing category which gives equal weight to each HS code. Second is to weight these codes by domestic output. Unfortunately, the output data is not available in the same disaggregated form as the HS codes. The avail- able output data is from the input-output table and includes 98 tradeable sectors. So the output weights include only the 98 sectors. Third method is to look at the distribution of imports. The codes whose import licensing category could not be identified are shown separately as "unknown." These unknown HS codes are distributed across the categories according to their shares in the total known items. Thus the presentation assumes that the items whose import licensing category are unknown are distributed in the same proportion as known items. Table 1 presents the structure of the licensing system under these three methods.  - 12 - TABLE 1: COMPARISON OF DIFEREN? MEASURES OF SNDIA' S !MPRT LICENSINUG STRCTUR DannedI LiaLted Percentage Restricted Permiesible OGL Canalized TOTAL Identified Percentages of BS Codes 46 26 19 9 100 71 Percentages of Output-weighted 56 24 11 9 100 76 HS codes Percentages of Imports 20 28 20 33 100 82 As expected, restricted licensing measured through HS codes is much more prevalent than measured through imports. This is because most items In the restricted lists are not imported. 33. The structure of imports and number of six digit HS codes (weighted by output) by licensing categories are given below In Tables 2 and 3. They are obtained by dividing imports (or HS codes) in different licensing categories to total identified imports (total number of output-weighted HS codes) in that subsector. The percentage of items whose import licensing category could not be identified is shown separately in the ounknown column. These unknown imports and HS Codes are distributed across the categories according to their shares in the total known items. This method seriously understates the restrictiveness of the import regime. First, by definition, all capital goods not specifically on the OGL list are restricted. Second, although raw matelials not on the restricted lists are supposed to be on OGL, the actual implenentation has not worked this way. Effectively, items not explicitly placed on OGL lists are restricted.  - 13 - TABLE 2: IMPORTS BY LICENSING CAEGORY (l7/88) (percentage ot total 1ports) of Imports in Gross Banned and Limited Denetic Restricted Permissible ML Canal zed Total Unknown/* OutRnt AGRICULTURE 42.0 2.0 46.0 10.0 100.0 23.0 0.92 ENERGY 0.0 0.0 0.0 100.0 100.0 6.0 44.06 MINERALS 0.0 0.0 8.0 91.0 100.0 16.0 23.98 KANUFACTURING 23.0 36.0 23.0 18.0 100.0 21.0 8.94 Food, Beverages and 34.0 0.0 0.0 65.0 100.0 26.0 2.51 Tobacco Textiles and Leather 40.0 38.0 22.0 0.0 100.0 22.0 0.61 Petroleum and Coal 0.0 42.0 20.0 38.0 100.0 91.0 6.48 Products Chemicals 6.0 26.0 40.0 29.0 100.0 13.0 12.14 Non-metallic Minerals 8.0 89.0 2.0 0.0 100.0 2.0 28.38 metals 16.0 26.0 10.0 47.0 100.0 12.0 13.97 Metal Products 0.0 94.0 6.0 0.0 100.0 28.0 4.41 Machinery/b 52.0 21.0 27.0 0.0 100.0 13.0 24.79 Electrical Appliances 2.0 56.0 40.0 2.0 100.0 5.0 16.21 and Electronics Transport Equipment 3.0 56.0 24.0 18.0 100.0 42.0 4.42 Others 12.0 17.0 47.0 24.0 100.0 47.0 10.15 TOTAL 20.0 28.0 20.0 33.0 100.0 18.0 7.17 /a These are given tor information only. In the calculations, they are distributed to other known categories, so that totals include all imports. /b Project imports are assumed to be restricted. 34. Table 2 presents the share of imports in different licensing categories by key subsectors. The energy subsector consists of crude oil and coal and lignite. Minerals include other non-processed raw materials. The last column in Table 2 gives the share of imports as a proportion of the gross output in that subsector. 35. In analyzing the restrictiveness of imports, two categories of relatively QR free imports have to be defined. First, OGL imports, while subject to actual user condition, can be imported without restriction on quantity. Their share in imports are abott 20X overall and 23X in manu- facturing. Second, items importable through export related licenses (REP); all items in limited permissible and canalized lists can be imported by REP  - 14 - licenses and freely traded in the domestic market. If the items importable by REP licenses are also treated as restriction-free, the proportion of commodities importable under non-restrictive categories increase to about 71%. This does not mean that 71% imports actually were brought in under OGL and REP. It only gives the proportion that can be imported under less restrictive categories. 36. The only other year that estimates of imports by licensing category was made was 1980/81; when the share of OGL imports was 6.5% in total imports and 7.9% of non-POL imports. These ratios have increased to 20% and 25% respec- tively, which support the hypothesis that the QR regime has marginally relaxed over the 1980s. There is also -me indirect evidence that the quantity of imports that are licensed have increased. The premia on imports have been very high in the 1960s and 1970s (Panchamukhi, 1978); increasing from 100% to 200% for many products. These rates have come down significantly in the 1980s (Pursell, 1988; Kishore, 1989). The average premia on REP licenses have been around 0-20% for most products. The decline can be attributed to currency devaluation after 1986, doubling of tariffs on imports of manufactured goods during the 1980s (Aksoy and Tang, 1991). By 1988, almost no product had consistently high import premia. 37. On the other hand, Table 2 also confirms that Indian import regime is still very restrictive. Imports on OGL are only 20% of all imports and constitute only 1.5% of total domestic output. In manufacturing, which is supposed to be relatively more liberalized, share of OGL imports in gross output is only 2%. The rest of imports (which are only 9% of domestic gross output in manufacturing) are further restricted through various licensing systems. The bulk of raw materials ln energy and minerals subsectors are canalized. Even in manufacturing, 33% of imports are still canalized by public sector importing agencies. 38. Table 2 also illustrates another problem of the import regime, namely its non-transparency. Despite best efforts, more than 30% of the HS codes and 20% of imports could not be classified. This lack of transparency creates tremendous problems for importers and producers in India. Every item that is not exactly specified in the Import document will be questioned by CCI&E, customs, and banks (which release foreign exchange) separately, leading to  - 15 - delays in imports and disruptions in the production activity. So simpli- fication and streamlining of the import policy according to international standards, such as HS codes, would by itself improve the system considerably. 39. Within the manufacturing sector there are large differences among industries in terms of restrictiveness of imports. The share of imports classified under OGL varies between 0% and almost 50%. Imports of raw materials such as chemicals, metals and food products are primarily canalized. Only three subsectors, chemicals (40%), machinery (27%) and electronics (40%) have significant amount of OGL imports. Since 1984, special policy packages have been introduced for petrochemicals, electronics and machinery, that contain partially liberalized import regimes. However, over the last two years, even these sectors have begun to experience more import restrictions. 40. Looking at the share of imports under different licensing categories gives a misleading picture of the nature of QRs in Indian economy. All the items that are not imported due to restrictions are excluded from the base, which only include the imported items. For example, if one Just looks at number of HS codes under QRs, then about 46% of the codes are in restricted category (as compared to 20% of imports) and only 19% are in under OGL. A more meaningful analysis would be to estimate the share of output protected by QRs. Table 3 presents the distribution of HS codes by licensing categories, weighted by gross outputs of each industry within the subsectors.j/ 5/ The distribution of HS codes in each industry is assumed to be the same as the distribution of output in that industry. Then these rates are weighted by gross output shares of each industry within that subsector. Within agriculture, for example, there are 22 activities (industries). In each of these activities, there is a distribution of HS codes by licensing categories. The distribution of HS codes for the whole agricultural sector is weighted by the gross outputs of each of the 22 activities. Other subsector ratios are similarly weighted by the gross outputs of industries within that subsector.  - 16 - TABLE 3: DISTRIUTION OF HS ORES B _LICENSIN6 P,=RV () (weighted by ao est c gross output) Number of Banned S Limited Industries Restrieted Permissible OIL Cnalized Total Unkn La AGRICULTURE 22 74.0 3.0 7.0 15.0 100.0 36.0 ENERGY 2 0.0 0.0 0.0 100.0 100.0 62.0 MINERALS 7 24.0 20.0 31.0 24.0 100.0 58.0 MANUFACMURING 67 52.0 28.0 12.0 8.0 100.0 21.0 Food, Beverages, and Tobacco 7 90.0 1.0 0.0 8.0 100.0 11.0 Leather and Textiles 10 55.0 43.0 2.0 0.0 100.0 8.0 Petroleum and Coal Products 3 24.0 26.0 8.0 42.0 100.0 41.0 Chemicals 9 17.0 43.0 32.0 8.0 100.0 32.0 Non-metallic Minerals 3 30.0 31.0 37.0 2.0 100.0 36.0 Metals 4 12.0 36.0 7.0 45.0 100.0 12.0 Metal Products 2 36.0 46.0 18.0 0.0 100.0 44.0 Machinery 10 42.0 24.0 34.0 0.0 100.0 31.0 Electrical Appliances 3 27.0 46.0 22.0 5.0 100.0 13.0 and Electronics Transport Equipment 6 40.0 30.0 30.0 0.0 100.0 45.0 Others 7 49.0 16.0 35.0 1.0 100.0 43.0 TOTAL 98 56.0 24.0 11.0 9.0 100.0 24.0 La Tnese are given tor information only. In the caiculations they are distributed to known categori`es,so~ that totals include all HS codes. 41. As shown above, the share of output-weighted HS codes under restricted licensing is much greater than the share of imports. For example, In manufacturing, while only 23% of imports are restricted, more than half of the output-weighted HS codes are in the restricted list. if the unknown category is also assumed to be restricted, then almost three quarters of output- weighted HS codes are in the most restrictive licensing category. 42. The distribution by manufacturing subsectors displays similar patterns. The proportion of HS codes that is in the restricted list for consumer goods such as food, textiles and other manufacturing subsectors is very high. In chemicals and metals, which are the key intermediates, the share of the restricted list is much lower. In chemicals the share of OGL (32%) is much higher while metals are basically protected through canalization. What is surprising is: despite attempts since 1978 to increase the items in OGL lists, the share of output-weighted HS codes under OGL is only 34% in  - 17 - machinery, 7K in metals and 32% in chemicals. If the unknown category Is also treated as restricted per Import Policy, the share of OGL for machinery comes down to about 24X, while the share of restricted list increases to 56X. How- ever the information in Table 3 is classified or interpreted, the import regime in India, even in subsectors which are supposedly liberalized, is still very restrictive. B. THE STRUCTURE OF TARIFFS 61 43. The relatively high level of import taxes (tariffs) in India makes them an important source of government revenue and a major influence on the pattern of development. Inport tariffs are the second largest source of government revenue (excise taxes are the largest) and have been the most rapidly growing major source of revenue during the 1980s. They provided about 35K of the Central Government and 26% of the combined Central and State Government revenue in 1987-88 and have increased 7.4K p.a. faster than nominal GDP during the 1980s. 44. India's customs tariff consist of three parts: (i) basic customs duties applied to the c.i.f. price of the import; (ii) an auxiliary duty, also applied to the c.i.f. price; and (iii) *additional* or "countervailing* (CVD) duties applied to the c.i.f. price plus the basic customs duty and auxiliary duty. Basic customs duties are mostly ad valorem, though there has been a recent increase in the number of specific rates. Auxiliary duties are always ad valorem. CVDs are usually equal to excise tax imposed on locally produced goods and are a mixture of ad valorem and specific rates. In November 1989 basic duties ranged from OX to 355K; auxiliary duties from OX to 45K and CVDs from OK to 50K. 45. Additional or countervailing duties are subject to MODVAT in exactly the same way central excise is. Buyers of imported goods can deduct CVDs paid on material inputs when calculating their excise obligation on outputs. Hence, 6/ This section is written Jointly with Umnuay Sae-Hau.  - 18 - where they are equal to excise taxes, CVDs do not provide protection to domestic producers Basic customs duties and auxiliary duty, together, provide an indication of potential tariff protection. 46. At first glance the tariff schedule appears very simple, with quite uniform basic customs duties for individual chapters. In practice the struc- ture of customs duties is complex because there are a very large number of Oexemptions' (zero or reduced rates) that must be traced separately for all three components, to ascertain the relevant rate for an individual commodity. These exemptions are so numerous that the table of scheduled tariff rates is virtually irrelevant. Almost all scheduled tariff rates are superseded by exemptions giving lower rates. Moreover, these exemption notification are frequently changed, making it even more difficult to ascertain the relevant customs duty rate for a particular good at any given time. The effective duty rate for aluminum ingots, for example, has changed seven .1mes since February 1987. Over 18 months the basic duty changed from 25X to 35% to 3,700 Rs/ton to, 60X to 2,000 Rs/ton to 1000 Rs/ton to 500 Rs/ton and finally to 2,500 Rs/ton. 47. Another complication is that exemption notifications do not always apply to commodities irrespective of end use. Frequently, they are user specific. In the analysis of nominal tariff rates, only the general exemptions have been taken into account. The magnitude of user specific exemptions can be only observed through comparing trade weighted tariff rates with actual collec- tions. Such exemptions were provided to the leather industry in 1979 and again in 1985. Materials, components and parts face lower duties when used in the electronics industry. So do components when used in the manufacture of fuel efficient cars. The extreme example is given by Pursell (1988) where stainless steel had eleven different tariff rates depending on the user. 48. There are two characteristics of the Indian tariff system that separate it from those of other LDCs. First, the absolute levels of tariffs (both nominal and collection rates) are very high. Second, again unlike many other countries, the tariff rates are increasing. Table 4 gives the comparative tariff collection rates for a group of comparable countries. As can be seen from the table, tariff collection rates are two to five times higher in India then in comparable countries.  - 19 - TABLE 4: CROSS-COUNTRY COMPARISON OF INDIA'S CUSTOMS TARIFF STRUCTURE INPORT DUTY IMPORT DUTY AS A PERCENTAGE OF AS A PERCENTAGE OF IMPORTS TAX REVENUE 1980 1987 1980 1987 India 29.7 61.9 24.8 34.8 Pakistan 24.6 24.7 34.8 38.8 Bangladesh 19.8 17.9 39.1 38.7 Indonesia 4.7 4.7 4.5 6.4 Thailand 10.1 11.1 21.8 20.6 Turkey /a 39.1 11.8 12.3 17.8 Brazil 16.0 8.7 8.7 2.5 Mexico /b 11.0 3.9 7.7 6.5 /a The figures for Turkey cover years 1979 and 1987. ab The figures for Mexico cover years 1980 and 1988. Sources: Except for Turkey, Indonesia and Pakistan, all revenue figures are from IMF: Government Finance Statistics Yearbook, 1989: and all import figures (CIF) are from UN: 1987 International Trade Statistics Yearbook, 1989. 49. Table 4 compares India's customs tariff collections with those of selected countries in 1980 and 1987 in terms of (i) percentages of imports which measure the protection given by the tariff system, and (ii) tax-revenue shares, which measure the importance of customs collections as a source of public revenue. In most developing countries, these represent the two main objectives of customs tariff. As a percentage of imports, India has by far the highest customs collections rate in the sample. At 61.9,r its tariff collection rate is more than twice Pakistan's 25X. Rates for the rest of the sample range from 4.7X in Indonesia (1986) to 17.9% in Bangladesh (1987). What is equally striking about India's customs collection rate is that it has been rising very rapidly over time - more than doubling in seven years (from 29.7% in 1980 to 61.9% in 1987). While part of the increase is due to relaxation of QRs, the collection rates have increased in almost all sectors. The only other country that has liberalized its trade regime but still has relatively high tariffs is Turkey. Part of this is caused by the liberalized import of luxury consumption goods which attract higher taxes and special extra levies. For the rest of the sample, actual tariff rates either remained little changed or, as in Brazil, Mexico, and Bangladesh, actually declined.  - 20 - Between 1980 and 1987, Brazil's collection rate dropped dramatically, from 16.0% to 7.1% while in Mexico it fell from 11% in 1980 to 3.9% in 1988. 50. As a share of tax revenues, India also ranks high in the sample. Import duties represented 34.8% of India's central government tax revenues in 1987. Although this is not as high as Pakistan's at 40.9% or Bangladesh's at 38.7%, it is far above those of other countries where rates range from 6.4% In Indonesia to 20.6% in Thailand. 51. When compared with the other countries in the sample, dispersion of India's customs tariff appears surprisingly low (Table 5). On the basis of nominal tariff rates, India has a coefficient of variation of 0.36, compared with 0.44 for Brazil and 0.81 for both Pakistan and Thailand. 52. The above comparison shows that India has an extremely high effective customs tariff rate and that the rate is high across the board as indicated by the relatively low dispersion of its tariff rates. Share of tariffs in total tax revenue, although not as high as Pakistan and Bangladesh, is very high. TABLE 5: CROSS-COUNTRY COMPARISON OF DISPERSION OF NOMINAL CUSTOMS TARIFF India Pakistan Thailand Brazil Nominal Tariff Rate 1988 1986 1985 1989 a. Mean 141.2 65.6 33.8 43.0 b. Standard Deviation 50.4 53.2 27.3 19.1 Coefficient of Variation, b/c 0.36 0.81 0.81 0.44 The Existing Tariff Structure 53. Table 6 presents India's customs tariff structure by subsectors. For each subsector, three tariff rates are given: (i) nominal, (ii) import- weighted, and (iii) collection rates. Tariff rates are further subdivided into *protective" and *total" tariff rates. Protective tariff is simply the sum of basic and auxiliary tariffs while total tariff equals protective plus countervailing (CVD) duty. Nominal tariff rates are based on the 1989-90 customs tariff schedules, while the collection and import data are for  - 21 - 1987/88. Since there has been only marginal changes in the nominal tariff rates this should not affect the analysis significantly.7/ TABLE 6: INDIA'S 1987/88 CUSTOMS TARIFF STRUCTURE TRADE-WEIGHTE DI~s T a wEI6 1E0-MNA UEARI O LIS 2 10tRATES customs~~~~~ _rtc FOO Poec- Ismorts Duties tive Total tive Total tive ]otL1 AGRICULTURE 4.3 1.6 87.8 90.3 37.5 39.7 21.7 22.0 ENERGY 14.6 14.2 92.6 115.1 60.7 60.7 55.8 56.8 HINERALS 1.9 0.8 99.5 103.6 24.3 27.5 20.6 23.1 MANUFACTURINOL& 79.2 83.4 119.5 146.9 83.3 105.4 57.1 65.9 Food Beverages and 3.9 4.1 135.1 147.1 80.7 89.8 60.3 61.2 Tobacco Textiles and Leather 1.1 1.4 136.1 158.4 119.8 170.4 44.1 72.6 Petroleum and Coal 4.9 2.7 127.0 174.1 13.1 32.2 12.6 32.6 Products Chemicals/b 13.7 20.4 118.3 156.4 94.0 123.7 70.5 87.8 Non-Metallic HineralsLe 9.5 0.8 126.3 171.0 118.3 168.3 64.5 79.7 metals 11.3 15.6 117.2 132.9 88.1 100.5 71.0 80.5 Metal Products 1.2 1.1 131.0 168.0 97.7 129.6 45.4 57.0 All Machinery 21.3 24.8 92.2 109.6 89.5 101.4 62.9 67.6 Electrical Appliances 3.6 3.9 105.1 144.6 97.1 131.1 52.1 63.0 and Electronics Transport Equipment 2.2 2.2 95.7 125.7 73.5 88.9 51.9 59.7 OthersLd 6.5 6.4 115.2 147.4 53.2 68.6 26.0 32.4 TOTALIa 100.0 100.0 116.0 141.2 72.3 85.5 4. 61.9 Based on reported total imports and collections, i.e., with no adjustments. Excluding tariff chapters 98 and 99 which have collection numbers but no recorded imports. Excluding gem imports for export production. Excluding tariff chapters 98 and 99 in Miscellaneous Manufacturing. 54. Nominal Tariff. The nominal tariff rates shown in the table are simple averages. India's average nominal protective tariff for 1989-90 is about 116X, with a standard deviation of 37X, while the average total tariff is about 142X, with a standard deviation of 51X. The difference between the two implies an average CVD rate of about 25X. 7/ More detailed tariff rates are presented in Appendix B.  - 22 - 55. The nominal protective tar1if rates are quite bunched together around 100l with raw material supplying industries such as agriculture, energy, minerals with tariffs of less than 10O, while the other sectors have pro- tective rates of more thar lOOX. The only exceptions are machinery and transport equipment. For total tariff rates, all rates except agriculture, are significantly above 100X. 56. Import-Weighted Tariffs. The protective and total tariff rates weighted by imports are lower than nominal rate by about 40 percentage points, with wide variation among different subsectors. The protective imported weighted tariff rates are 72X for the whole economy and 83X for the manufacturing sector. Total tariff rates are 86X and 105X respectively. The variance among different subsectors are greater for import weighted tariffs than the nominal ones. Goods that are allowed to be imported have significantly lower tariffs; indicating that imports are restricted to products that are not available in India. 57. Actual Collection Rates. Actual collection rates are less than one half of the legal nominal rates and are on average 30 percentage points lower than trade weighted rates. The difference between trade weighted and actual pro- tection rates are due to specific exemptions (specific to users) and duty free imports for export production. For example, in non-metallic minerals sub- sector duty free imports of rough diamonds for processing into exports are excluded. In other subsectors, it is not possible to separate duty free imports for exports. However, total duty free imports (excluding gems) were 4.4X total imports in 1987/88. The total tariff collection rate on manu- factured goods imports excluding duty-free imports for exports was 80X, compared to 65.9X for all manufactured goods imports. 58. The collection rates although lower than nominal or trade weighted rates, are also very high with protective tariffs of 55X and total tariff of 62Z. Moreover, protective tariff rates are highest in three subsectors which have a strong cost impact on the rest of the economy; metals (71S), chemicals (70X) and machinery (63X). Furthermore, these the subsectors also account for about 45X of all imports. The sectoral average tariff rates given in Table 6 mask the very high rates in some key industries within sectors. For example, protective tariff rates are 106X and 122X, respectively, in organic heavy  - 23 - chemicals and other chemical industries, compared to the average of 70X for chemicals as a group. In non-ferrous metals, the protective tariff col- lection rate is 82X. For most industrial machinery imports, the rate was 75X. These rates can only be justified as trying to earn maximum revenue from a very small import base. Very few of the firms in these sectors require these levels of protection for their viability (Aksoy and Ettori, 1991). Evolution of Tariff and Excise Tax Collection Rates 59. During the 1980s, the tariff collection rates for most countries have been decreasing as they move away from overvalued exchange rates and liberalize their trade regimes. In India, the opposite has happened. Historically, to contain balance of payments crises, tariff rates were increased instead of adjusting the exchange rate. However, these rates were not reduced when the exchange rate was eventually adjusted. Thus with every foreign exchange crisis, the average tariff collection rate has ratcheted upward to a higher plateau. Figure 1 shows tariff collection rates since 1960/61 for total and manufacturing imports (total imports minus food, fuels (POL), and gems). FIG. 1 TARIFF COLLECTION RATES 80. 70 60. 50- 40- 340 10 I I , . . . . . , . . . . . . . . . . . . 62 64 66 68 70 7274 7f6 78 80 82 84 86 88 -|TOTAL IMPORTS .. MANUPACTURING IMPORTS!  - 24