Policy, Reserch, and External Affairs WORKING PAPERS International Trade international Economics Department The World Bank March 1991 WPS 610 Can Preshipment Inspection Offset Noncompetitive Pricing of Developing Countries' Imports? The Evidence from Madagascar Alexander J. Yeats Many developing countries pay preshipment inspection firms well to verify that imports (and sometimes exports) meet quality and quantity standards and that prices are within established norms. But preshipment inspection failed to reduce the exces- sive import prices Madagascar was paying (particularly for chemicals and basic manufactures), possibly as the result of false invoicing by Madagascar importers and industrial country ex- porters. 'TIhe N)! i. y Researchh and 1 Icrr a Aff:rs Compnicx d5tnrvte s I'Ri W'orking Papers to dissriinar e the find ungs of wo:k i progress and to ncourage he ex c"anrge of ideas among flank staff and a1' oAheni intcrested un development issues. These papers carry the names of the authors, nrect clon'v heir * icw%, and should he used and cited accordingly 'TIhe findings, knterprtuuons, and conclusions are the authon' own I hc) shs>'d not he atnl teo the World BHank. it Board of l)irrectom ,its managencnt, or any of ius merrher counries PONioY, owoh, adExml At Ir International Trod* WPS 610 This paper- a ploduct of the Interamaon Trade Division, I rematon Economics Deparlment - is part of a larger effort in PRE to improve developing countries' ability to make more effective use of their financial resources in the procurement of imports vital to industrialization and growth. Copies ame available free from the World Bank, 1818 H Street NW, Washington DC 20433. Please contact Jean Jacobson, room S7-037, extension 33710 (31 pages). Many developing countries use preshipment suggest that Madagascar paid considerably inspection (PSI) firms to counter the adverse higher prices than other developing and indus- effects on their foreign trad. .f certain pricing trial countries both before and after PSI was and business practices. These firns may also adopted. perform some national customs functions, but Lheir key responsibility is normally to verify that In other words, preshipment inspection imports (and sometimes exports) meet quality failed to reduce Madagascar's import prices to and quantity standards and that prices are within the level of those paid (on average) by other established norms. importers. Extreme prices (150 percent or more above average) occur for all typcs of goods Developing countries make substantial imported by Madagascar but are clustered in payments for PSI - charges appear to average chemicals (SITC 5) and basic manufactures about I percent of the value of the goods in- (SITC 6). spected - but have undertaken no comprehen- sive cost-benefit studies of PSI. Evidence suggests that collaborative false invoicing by Madagascar importers and indus- Using data from Madagascar's experience, trial country exporters is one reason for the Yeats analyzes the impact of PSI on excessive prices both before and after adoption Madagascar's relative impon prices. The results of PSI. The PRE Working Paper Series disseminates the findings of work under way in the Bank's Policy, Research, and Extemral AffairsComplex. An objectiveof the series is to getthese fmdings out quickly, even if presentadons are less than fully polished. l The findings, interpretations, and conclusions in these papers do not necessarily represent official Bank policy. Produced by the PRE Dissemination Center Can Preshipment Inspection Offset Noncompetitive Pricing of Developing Countries' Imports? The Evidence from Madagascar by Alexander J. Yeats Table of Contents I. Introduction II. Scope and Methodology of the Study 4 III. The Industry Pattern of Relative Prices 14 IV. Summary and Conclusions 19 References 22 Appendix 1: Analysis of Madagascar's Relative Import Prices 24 and Expenditure -- Effects for Five-Digit SITC Iron and Steel Products, 1979-82 and 1984-88 Appendix 2: High and Low Tariff Products for Tests of False 29 Invoicing by Industrial Country Exporters and Countries Covered by Madagascar's PSI Program Can Preshipment Inspection Offset Noncompetitive Pricing of Developing Countries' Imports: The Evidence from Madagascar Alexander J. Yeatsl I. Introduction While numerous theoretical and empirical studies have examined the positive contribution that trade makes to developing countries, industrialization and growth, several recent investigations suggest that the conditions under which some trade occurs may restrict its positive effects. For example, a World Bank analysis of European countries' prices for iron and steel goods showed former African coloriies pay 15 to 25 percent more, on average, for imports than other industrial or developing countries, and that these adverse price differentials persisted over (at least) the last three decades.2 In addition, previous analyses of discrepancies in partner-country trade statistics provide evidence on the existence of illegal practices such as smuggling and false invoicing to evade tariffs or other restrictions, or to effect capitPl flight.3 Cases have lPrincipal Economist, International Economics Department, The a.rld Bank, Washington 20433. The views expressed in this paper need not reflect those of the World Bank, its management, or its member countries. 2See Yeats (1990) for details. Factors which appeared to be responsible for the adverse price differences include the small size of the African countries relative to other importers, and the lack of aggressive competition by exporters. In a related study Hufbauer and O'Neill (1972) also show that small countries typically pay more for imports. Avramovic (1978) determined that market imperfections, as well as a lack of finance and countervailing power, result in developing countries generally receiving less than developed countries for exports. 3For example, Bhagwati (1967) and Sheikh (1974) use partner-country trade data to show how Indian and Pakistani exporters inflate invoices to illegally secure export subsidies. Simkin (1970) uses the same approach to assess the level of smuggling and noncompliance with international commodity agreements in Afriea. See Ely (1961) and Morgenstern (1963) for a general discussion of fa-tors causing discrepancies in partner-country trade data. 2 also been cited (see Helleiner, 1978 or Edwards, 1972) where transnational corporation practices, international cartels pooling and allocation of patents, trademarks and copyrights, allocation of territorial markets and other restrictive business practices reduced competition in import and export markets and developing countrias' gains from trade. In an attempt to offset the detrimental effects of such practices, a growing number of developing countries have engaged the services of preshipment inspection (PSI) firms to verify that the quality and quantity of goods shipped meets contractual standards and that the prices charged are within "reasonable" norms. 4 Considerable costs are involved as the United States International Trade Commission (1987) estimates the PSI companies, charges average about three 4A tabulation by the U.S. International Trade Commission (1987, p. 1-4) indicates the following countries were using pre-shipment inspection services as of December 1986 (starting date in parentheses): Angola (1980), Bolivia (1986), Burundi (1978), Congo (1987), Ecuador (1985), Equatorial Guinea (1983), Ghana (1971), Guatemala (1986), Guinea (1986), Haiti (1983), Indonesia (1985), Cote d'Ivoire (1975), Jamaica (1986), Kenya (1972), Liberia (1986), Madagascar (1903), Mexico (1985), Nigeria (1984), Paraguay (1983), Philippines (1986), Rwanda (1977), Suriname (1982), Tanzania (1972), Uganda (1982), Venezuela (1986), Zaire (1968), and Zambia (1978). The Societe General de Surveillance (SGS) was the PSI firm being employed exclusively by all but four (Congo, Guinea, Nigeria and Venezuela) of these countries. 3 quarter, to one percent of the value of goods inspected.5 Proponents argue (see Mowbray (1988), Dornbusch (1987), or Societe General de Surveillance (1989) that the service is cost effective since preshipment inspection prevents price gouging by sellers and false invoicing to avoid tariff and tax liabilities or effect capital flight, combats shipment of substandard goods or items that otherwise fail to meet contractual requirements, and can be used to verify that excessive freight charges are not levied on imports (see Societe General de Surveillance (1988)). In spite of the importance of these problems, and the claims concerning the utility of PSI for dealing with them, there appear to have been no comprehensive anElyses aimed at evaluating the performance and results of reshipment inspection. The present study provides some relevant information by analyzing Madagascar's relative import prxces befcre and after PSI requirements were adopted. In particular, an attempt is made to determine if Madagascar paid "inflated" prices for some goods and, if so, how effect.v- preshipment inspection was in countering this problem. Also, statistics relating SThe PSI companies focus almost exclusively on imports, although they have been employed for some export products, especially when subsidies and other incentives are offered. While the actual services performed differ from country to country, the normal PSI contract covers the following 14 basic points: (1) the purpose of the contract; (2) the nature and scope of the inspection services to be rendered; (3) obligations regarding comparison of prices; (4) obligations of the contracting government; (5) identification of the goods subject to inspection and those to be exempt; (6) special procedures regarding inspections of goods from certain countries; (7) exempt transactions; (8) reporting requirements; (9) obligations of the inspection company and vendora; (10) fees and other charges; (11) method of payment; (12) liability; (13) resolution of disputes between the contractor and government; and (14) term of the contract. Regarding point (6), exports from the (former) socialist countries of Europe and Asia and other developing countries are generally exempt from inspection, although some Eastern European countries are covered by Madagascar's contract. 4 to false invoicir- are analyzed to determine if PSI was effective in combatting capital flight or customs duty avoidance. TI. Scope and Methodology of the Study Madagascar was chosen as the subject of the present investigacion for several reasons, including the fact that the preshipment inspection progr with SGS was adopted in 1983 and the required data were available to assess the effects of the program for its first five years of operation.6 Madagascar was also selected due to the comprehensive nature of its preshipment inspection requirements--inspections are performed on v_.rtuall- all imports of general merchandise, equipment and materials, most types of machinery (especially that destined for "infrastructure, industrial, and agro-industrial projects.-)' Imports valued under 4 million Malagasy francs (about USS5,900 at 1986 average rates of exchange) are, however, exempted from inspection. The choice of Madagascar was based on claims concerning the program's effectiveness and 6See USITC (1987, pp. 3-81 to 3-86) for a general discussion of the features of Madagascar's program which began on 1 June 1983. Appendix Table 3 lists countries where PSI of exports to Mada-;.,dr is required. The cost of the program is estimated by USITC to be about 1.4 percent of the f.o.b. value of each import license. 7Products exempted include: gold; precious stones; works of art; explosives and fireworks; munitions, weapons and instruments of war; live animals; fresh, frozen, or refrigerated fish; eggs; fresh, refrigerated, or frozen meat; fresh, refritsrated, or frozen fruit and vegetables; salvage metals; personal belongings and household goods, includinc, one used vehicle; current newspapers and periodicals; imports through the mail; gifts; supplies for diplomatic and consular missions; and supplies for agencies of the United Nations that are imported for their own needs. Inr.tead of listing countries exempted from inspections, the contract lists coun;ries where inspections are to be performed (This list can be found in Appendix Table 3.) Another exemption is that the price comparison is not required for raw petroleum and petroleum prodacts delivered in bulk. Only quantity and price inspections are required for pharmaceutical products, dyes, paints, insecticides, pesticides and fungicides, special chemical products, cosmetics, wines (except in bulk) and brand-name spirits. Special chemical products are defined as any chemical product produced exclusively by a given manufacturer with a confidential or protected trademark. 5 savings--the General Director of Foreign Trade for Madagascar has stated that inspections save the count.y a minimum of FMG 500 million (about US$740,000) annually although no indication was given as to . w this estima-e was derived.8 Finally, Madagascar's imports largely originate in a relatively few industrial countries (with France by far the most important, see Table 1) which, with the exception of the United States, compile the detailed value and quantity trade statistics required for this analysis. Table 1: The Origin of Madagascar's Imports Major Product Groups, 1979-86 Share of Madagascar's imports originating in (X) Att iTmports European Community (10) Product Category (SITC) Year CS miIlions) Total France Germany Italy U.K. Japan EFTA USA Ali Goods (0 to 9) 1986 373.6 47.8 31.3 6.7 2.8 3.2 6.5 2.6 10.7 1985 465.1 46.6 29.4 5.6 2.0 4.6 2.7 1.7 16.2 1983 411.5 51.6 35.5 4.3 3.9 2.6 4.1 1.8 8.6 1981 473.0 61.9 37.6 10.0 5.1 2.7 3.1 6.4 4.4 1979 698.4 48.5 30.0 8.6 3.8 1.9 5.5 2.8 10.0 Manufactures (6 to 8 Less 68) 1986 176.5 70.1 44.9 10.2 4.3 6.5 9.1 2.7 9.9 1985 252.4 58.S 36.3 7.4 3.1 7.0 4.6 1.5 26.4 1983 193.1 72.0 49.0 6.7 5.9 5.3 3.2 1.9 11.1 1981 293.7 66.4 40.7 12.0 5.7 3.2 4.5 9.2 3.7 1979 377.3 56.9 :S.! 11.0 .. 1 3.1 9.4 3.5 16.5 ChemicaLs (5) 1986 47,0 73.8 52.1 .S 3.8 0.6 1.5 7.4 4.9 1985 49.0 83.0 57.6 .0 2.0 6.3 1.6 5.9 3.2 1983 41.6 82.5 62.7 7.2 3.1 1.0 3.1 7.5 1.4 1Q81 45.5 84.2 54.5 14.3 4.6 3.5 C.9 5.9 2.2 1979 76.8 77.0 44.4 20.2 4.3 0.9 0.8 5.6 1.7 Foods (0+1+22+4) 1986 56.9 19.5 12.5 1.0 0.7 0.2 13.5 0.5 33.4 1985 58.5 25.0 9.9 0.9 0.2 1.2 0.3 9.6 1983 82.3 25.0 12.0 0.4 3.0 0.1 11.4 0.2 14.5 1981 66.6 52.3 33.5 5.0 0.6 1.7 0.9 6.4 1979 105.5 34.6 27.1 0.6 0.2 0.9 - 0.3 4.2 Souree: Madagascar's reported imports as recorded in United Nations Series 0 Trade Tapes. 8Published statements from Midi Madaaasikara, February 16, 1987. 6 As a first step, annual value and quan..ity data were drawn from UN Series D trade tapes for French, German, Italian and Japanese exports to Madagascar over the 1979-1988 period and unit values were computed. This provided a "benchmark" on average prices (unit values) paid for the four years (1979-82) prior to the adoption of PSI ai well as five full years (1984-1988) afterwards. In general, the data were drawn at the very detailed five-digit SITC (Revision 2) level although some four-digit products were included in the Italian, German and Japanese statistics when more disaggregated data were not available.9 Value and quantity data were also drawn for exports of these products to other develop g and industrial countries so Madagascar's relative import prices could be 9The selection generally included every five-digit product exported to Madagascar over most of the 1979-88 period for which both quantity and value data were available. Certain products, such as those traded irregularly, or which clearly had diverse characteristics (i.e., f4ve-digit items with "not elsewhere classified" or "not elsewhere specified" headings) were excluded. Data on United States exports were not used since this country generally did not compile quantity information required for computation of unit values. A poLnt to note concerning the unit value information is that quality or product-mix variations may make price comparisons unreliable for some soecific products, but their influence should cancel out in the large number of products included in this study (i.e., there is no reason to believe that Madagascar is generally a purchaser of relatively high- or low-priced goods). See Appendix 1 for separate price comparisons for goods with homogenoub characteristics. 7 computed.13 This procedure produced free-cn-board export prices for srmilar goods shipped from the four industrial countries to Madagascar and other destinations. Table 2 provides summary statistics on Madagascar's relative pri:es (expressed as a percentage) for each year over 1979-88. Both sri;ple and trade- weighted (by Madagascar's import values) average prices are shown for shipments from France, Germany, Italy and Japan, along with similar statistics for these 10Several modifications were made in the comparator country groups. Since some Sub-Saharan African countrie~s use pre-shipment inspection, and a study by seats (1990b) showed their imr-.. prices were not representative of those paid by other countries, they were excluded from the developing country group. Also, Greece, Spain and Portugal were included in the industrial country group even though they are categorized as "developing" ir. some World Bank classLfications. These tabulations permitted calculation of Madagascar's relative import price (Rm) for each five-digit product i (1) Rm Vmi Qmi Vc; where Vmi and Vc* are the free-on-board value of imports by Madagascar and the comparator group (i.e., industrial. or other developing countries), respectively, and Q is the corresponding quantity. 8 Tabte 2: Relative Inport Prices Paid by Madagascar ancd Other DeveLoping or Industrial Countries from Selected Exporters, 1979-88 (percentage) Exporter Price Coeparator/Average 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 France Industrial Countries - WeightedC 34.7 42.1 38.5 44.0 37.6 27.8 29.0 34.3 56.5 23.3 Simiple 78.6 84.8 119.2 93.1 63.7 64.6 35.2 94.3 106.6 49.4 France Developing Countriesb - WeightedC 36.1 28.4 23.3 29.6 31.8 15.0 13.9 22.4 40.3 18.1 * Simiple 58.9 48.8 69.1 72.6 6.7 36.7 8.2 56.3 45.8 33.3 Germany Industrial Countries - Weightedc 71.1 7S.1 41.7 120.3 154.9 79.9 88.2 81.9 76.7 73.2 - Simpte 110.4 91.3 75.4 115.9 130.4 103.6 98.3 86.7 106.5 72.1 Germany Developing Countriesb Weightede 38.6 48.9 27.2 53.8 72.6 48.6 49.5 51.3 43.9 37.8 Simple 64.0 61.9 53.8 79.2 131.1 75.2 69.5 57.1 88.6 48.2 Italy Industrial Countries Weighte.4; 59.0 129.4 36.2 111.6 72.7 47.6 45.7 40.1 54.4 a Simple 38.8 30.7 28.7 113.0 99.9 36.8 14.2 32.6 37.6 a Italy Developing Countriesb Weight'dc 46.4 98.7 37.4 80.4 -3.6 43.4 39.6 39.6 46.0 A Simpte 42.6 21.5 35.0 111.3 82.9 42.9 17.5 32.9 38.2 a Japan Industrial Countries - Weightedc 71.6 32.7 172.3 85.0 133.7 75.0 62.2 107.3 36.3 68.6 -Simple 62.3 95.7 202.7 89.4 163.4 134.3 112.6 248.2 192.5 150.4 Japan Developing Countriesb Weightedc 93.9 63.2 163. 105.0 107.6 53.7 66.0 102.9 68.2 75.0 Simple 116.1 142.7 199.9 129.5 173.8 165 7 128.2 210.9 218.5 186.4 ALL ABOVE Industrial Countries - Weightedc 45.6 56.0 46.9 66.3 67.6 42.9 42.8 41.2 56.4 29.8a Simple 78.2 77.0 95.6 101.7 93.6 74.1 52.6 92.5 99.8 67.9a ALL ABOVE Developing Countriesb - Weightedc 40.9 40.5 33.7 43.0 47.5 29.8 27.1 39.4 40.6 22.9a - Simple 61.8 55.2 64.4 87.0 77.7 60.0 35.3 64.6 70.3 55.38 a Price comparisons for Italy could not be made for 1988 due to a lack of quantity information on exports. b For definition of developing countries see footnote 10. c Weighed by Madagascar's import values. Source: United Nations Series D Trade Tapes. 9 four countries combined. As ouch, Table 2 indicates the extent to which Madagascar's annual import prices exceeded other countries (positive numbers show such a percentage price premium) while any negative values (a discount) would show the percentage by which they were lower. Table 2 indicates Madagascar always paid a substantial price premium for imports--irrespective of whether comparisons are made with induetri 1 or other developing countries. The lowest observation, based on 1985 French prices, indicates a Madagascar premium of 8 percent over similar goods shipped to developing countries, but in most years the relative prices are far higher. In fact, several comparisons, such as those for Germany's or Japan's 1582 and 1983 exports show Madagascar's import prices were more than dcuble those of other developing countries, while Japan's 1981 (trade-weighted) prices were 200 percent higher. liSpecifically, Madagascar's unweighted average relative import price was derived from: Vmi Qci x (2) Um r Qmi V N where the V and Q terms are defined in equation (1) and N is the number of five- digit goods shipped. The trade-weighted average (Wi) is defined by: Vmi Vmi Qci Vmt Q^ Vci (3) Wm = N where Vt represents Madagascar's total imports of the five-digit products. 10 Two other poLnts ct importance are evident from Table 2. First, the trade- wei.ghted price rp.atives are z_rslstently lower than the simple averages (see *lbo Table 3) w'.ch Lp' es Madagascar pays lower relatLve prices on larger purchases. Sin-e Yeats ( 9-Cb) and Hufbauer and O'Neill (9'372) report similar findings it suggests prccedures such as bulking domestic orders, or making joint purchases with other countries may result in lower prices. Second, the average price relatives are almost always higher when industrial countries are the comparator group--a po;nt that indicates developing countries typically pay higher prices for similar products (see Yeats (1990b, Table 4 on page 10 for related findings). (This is not so often the case for Japan as for the European countries.) Table 3 is addressed to the key question of this investigation--does the evidence suggest that presh.pment inspection improved (lowered) Madagascar's relative import prLces.12 Specifically, the table shows Madagascar's average (weighted and unweLghted) price relatives before and after PSI was adopted. In the 1984-88 period (when PSI was required), Madagascar's (trade-weighted) import prices averaged 43 percent more than those of industrial countries (unweighted average prices were about 77 percent higher) and were about one-third higher than other developing countries. Although these averages are down somewhat from the pre-PSI period, the differences are not statistically significant at the 95 12It should again be stressed that the present tests only relate to the impact of PSI on import prices--they do not show how well (or poorly) PSI fulfills other objectives such as speeding goods through customs, insuring that goods meet contractual standards, improving customs' revenue collection, or ensuring that quantities shipped are correct, etc. It would be difficult to undertake such evaluations using United Nations trade data, rather an on-site assessment of PSI is required. 11 Table 3: Average Price Premia Paid by Madagascar for Imports from France, Germany, Italy and Japan Before and After the Adoption of Pre- Shipment Inspection Requirements Intervals Before and After PSI Reauirements Price Comparator/Average Before (1979-82) After (1984-88)b Industrial Countriesa - Weighted 53.7 42.6 (45.8) - Simple 88.1 77.4 (79.8)c Developing Countries - Weightdd 39.5 32.0 (34.2) - Simple 67.1 57.1 (57.6,c - France, Germany, Italy and Japan, Italy excluded in 1988. b Figure. in parentheses are for 1984-87. The other 1984-88 averages do not include 1988 Italian prices. See the notes to Table 2. c Not significantly different from the 1979-82 average at the 95 percent confidence level. Significance teats were not run on the weighted averages. percent level (i.e., the data do not indicate that preshipment inspection led to statistically significant price reductions). A further point is that the improvement in 1984-88 relative prices reflected in Table 3 would be expected due to economic reforms adopted by Madagascar during this period (see footnote 13). While Madagascar's average relative import prices did not improve under PSI, there may have been some influence on their overall distribution. Specifically, it is possible that the averages basically remained the same in the pre- and post-PSI periods but the frequency distribution changed. Such could 12 be the case if PSI reduced cases where extreme price differences (say 100 percent or more) occurred, or if the variance in prLces about the mean were lowered.13 Table 4 summarizes information on the distribution of Madagascar's relative import prices during 1979-82 and 1984-88. Shown here are decile values for import price relatives from France, as well as for Germany, Italy and Japan combined. That is, the table entries corresponding to any given Di indicates that lO percent of all Madagascar's price relatives fell below the value shown while (100-10) percent were higher.14 As Table 4 shows, PSI had little or no influence on the frequency of extreme relative price differences. In both the 1979-82 and 1984-88 periods, 10 percent of all French exports to Madagascar had unit prices of 150 percent or more above those paid by industrial countries--a pattern very similar to that 13A potential limitation of the approach employed in this paper is that all improvements in relative import prices are attributed to the adoption of PSI, when they in fact may be due to other factors which lowered incentives for false invoicing and capital flight. According to the World Bank (1989), Madagascar undertook several major reforms in the 1984-88 period that should have, on balance, lowered its relative import prices. In 1987 and 1988, a market- determined trade and foreign exchange regime was adopted that included the elimination of quantitative import restrictions and also simplified (reduced) tariffs. While economic growth sharply deteriorated between 1980 and 1982, financial stabilization and a limited expansion were achieved from 1983 through 1988. These developments should have improved Madagascar's "credit worthiness" and reduced finance and insurance costs for imports. Since the latter are reflected in exporter's f.o.b. unit values they should have reduced 1984-88 relative import prices. Evidence also suggests that the black market exchange rate dropped in 1984-88, a development that should have had a positive impact on import prices. See Pryor (1988, Table G-2, p. 17) for estimates of the black market premium. 14As an example, Table 4 shows (see the entry corresponding to D8) 20 percent of French 1979-82 exports to Madagascar had unit values 116.9 percent or more higher than similar goods exported to industrial countries. In 1984- 88, 20 percent of these shipments had unit values that were 112.3 percent higher. Table 4 indicates the other industrial countries' (Germany, Italy and Japan) distribution was even more skewed toward high relative unit values--in 1979-82 20 percent of Madagascar's import values exceeded those for industrial countries by 140.8 percent or more. 13 Table 4: Analysis of Relative Irport Prices Paid by Madagascar Befcre anq Atter Adoption of Preshipment Inspection Requirements (Data correspond to reiative prices at deci.e imits) France Germar, a Italy and pa_an Cmroarator Industrial Comparator::eveioDp Comparator:Industnai Con atoror Deve o inq DeciLe 1979-82 1984-88 1979-82 1984-88 1979-82 1984-88 1979-82 1984-88 01 -29.1 -43.7 -38.3 50.7 -39.5 -48.9 -40.8 45.4 02 -9.6 -24.4 -20.0 -39.8 -20.5 -30.0 -18.6 -26.6 03 6.3 -4.3 -3.3 -21.8 *5.2 -17.2 -3.3 -8.7 04 20.4 10.7 8.2 *6.0 11.8 3.1 12.0 8.1 05 36.9 26.1 19.2 9.7 31.0 21.5 28.6 22.3 06 51.6 42.0 33.7 25.3 59.3 40.5 51.8 41.7 D7 78.4 65.8 52.1 44.2 92.1 79.6 79.7 68.3 08 116.9 112.3 77.9 69.8 140.8 126.5 112.5 112.5 09 151.7 151.3 137.3 120.7 152.2 151.8 151.6 151.6 Note: The deciLe Limits show the proportion of Madagascar's relative import prices that tie above or beLow certain values. For instance, during the 1979-82 period exactly half (50 percent) of Madagascar's import prices from France were 36.9 percent higher than those charged industriaL countries while half were lower. Similarly, 30 percent of Madagascar's prices were 78.4 percent higher. for Germany, Italy and Japan's prices. The shifts that occurred were in the lower deciles (i.e., the entries corresponding to D1 through 03 ranges). 15 The table also shows that some reduction cccurred in the median relative import prices--entries that correspond to the D5 values--due tc an increase in the number of products with an apparent price discount. However, tests on the mean prices (previously cited) indicate the reduction was not significant. 15A Chi-square test indicates that the 1979-82 distribution of relative import prices was significantly different from the 1984-88 distribution at the 9S percent confidence level. However, as Table 4 shows this is due to shifts in the middle and lower decile ranges and not to reduction in extreme adverse price relatives, i.e., those above the D9 limit. 14 III. The Industry Pattern of Relative Prices Several important questions concern the product groups in which the extreme price relatives occu --in particular, are they clustered in sectors where preshipment inspection is not required (see footnote 7 for information on this point), do they flag industries where PSI is less effective (possibly due to complex or secret pricing practices), are they in less-competitive sectors where monopoly pricing is a factor, or is there evidence they result from collaborative false invoicing by buyers and sellers. Using data on French exports, Table 5 allows examination of the distribution of these extreme price relatives before and after PSI was adopted. The table shows the percentage of extremes in each major one- and two-digit SITC group and also indicates Madagascar's overall import price relativity. 6 Finally, the table also gives the value of Madagascar's imports in each product group. Table 5 shows that the extreme price relatives are heavily clustered within chemicals (SITC 5) and basic manufactures (SITC 6), and that the introduction of PSI did little to change their distribution or frequency of occurrence. During 1984-88, 36 percent of all extremes occurred for chemicals (up 2 points from 1979-82) with over two-fifths of these observations in inorganic chemicals 16In 1979-82 there were 94 five-digit SITC products which had "extreme" price relatives (150 percent or more difference between Madagascar's and other countries' prices) while there were 111 such observations during 1984-88. The expenditure discount or premium measure (Ed/p) reported in Table 5 shows the percentage difference between Madagascar's actual and potential expenditure if the same quantity of imports were purchased under other countries' prices, (4) Ed/p ( mqm _ 1 ) x 100 Zpdqm A positive value shows the percentage "excess payment" associated with Madagascar's higher import prices. Table 5: the D'-.rr rot I ot Mad.qga%car s Irmport Pr ice Prvrfiriurii or Dbicounts arnd Extreme ReLative lImiport Prices Vatue ot irernch tlports Madagascai s expenditure discount Percentage of emtr rmw pr 1 r I tu t v vt _IS --r) _ __ or premium (,)a values IrV r.,, 4) SITC _DesrrTtron_ 1979 82 1984-88 1979 to 1982 1984 to 1988 1979 to 1982 1984 to 1988 0 FOOD AND LIVE ANIMALS 62.2 40.7 -35.8 6.2 0.0 0-9 2 CRUDE MATERIALS FXCEPT FUELS 16.8 22.5 1.3 8.2 i.2 356 3 MINERAL FUELS 22.7 4.7 44.5 42.1 1.1 O0C 4 ANIMAL AND VEGETABLE OILS 2.1 0.8 36.4 77.8 0.0 356 S CHEMICALS 111.8 112.7 42.3 27.1 34.0 36-1 of wshich: 51 Organli Cherrriial 8.5 6.7 63.9 43.1 9.6 9 '. 52 ltiorgani Chumi(dis 12.5 5.5 38.0 3.4 13.7 1S4 54 MVedllnal PJrdJcts 52.9 38.2 51.0 54.1 1.1 U 9 58 PdA.TIc M,t ir ,tr-. 9.4 ?9 1 17.8 14.3 t.3 4 59 Cheurnralk or.. 22.6 21.8 22.9 26.0 4.3 653 6 BASIC MANUiACtLJR[S 155.7 125.1 21.9 14.5 34.1 53 of wh ch: 62 R~ubh(*r 14,0rt.,1 tLr ,r, ?1.6 18.7 45.3 7.2 1.1 64 Pdaef arid Mr.ui.fa,r rule-S 9.5 10.5 26.8 33.9 5.3 65 textite Ya,n .Irsl Fabrits 15.3 16.1 -11.3 15.8 4.3 7.2 66 N-rmrvtat Mir ii-, Mirrufic ture 9.3 11.2 20.8 1.0 4.3 it. 67 Iron ndd Stu-I 52.0 35.3 21.2 19.9 2.1 1-8 68 N.,rrfer ior Mutal, . 6.4 4.3 43.8 26.8 7.4 8.1 69 M-tal Mar..t.r. tO. r-, oi. 39.1 27.9 6.8 -3.6 9.6 5.4 7 MACHINIRt ANlD IkANSi (OT lQLU8PMLNi 311.9 264. 3 19.4 18.2 14.9 13.5 Ot whtItlr 71 Power (.eneritiq Ir Equipment 7.0 27.4 17.4 21.2 0.0 1.8 72 M,thirner f ! .p ral lniiutrles 104.5 '5. i 21.3 3.2 6.4 3.6 74 Gr-neral I rx,- I i ia Ma(hrnety 50.? 45.1 22.8 -41.0 2.1 1 8 75 ot rre Ma( h rre arKi EqkJurpnent 6.8 13.0 20.5 1.8 2.1 0. 9 76 7 1i e r rwmnn, rv.1 f r qui prnent 25.4 16.6 29.5 29.0 1.1 0. 0 77 Hri ( ,,i Murhir Mh y ne. 36 .1 44.9 9.8 8.6 3.2 5.4 8 MISC. MANkJFAi Ti f. D i.tljODS 36.0 38.2 19.4 16.9 12.7 9 0 Of Wtir Ui 82 Forn t LJr 1.2 1.0 -4.6 26.8 2.1 84 Clkthing 1.7 1.4 -34.2 -2.7 2.1 t.8 87 Prer_a,us Irnr. rwkrt% 9.9 11.4 51.2 12.6 1.1 ' 8 89 M .c. Manfta, turn. 16.7 16 .4 31.2 -30.4 7.4 S.4 a Defined as the .a Tri. 1r.-ymnet made by Madagascar tfr itemrs in the group divided by the payment required if Madagascar t-*a iridaIr a lt courtry prices wIth the re..tt evpre,sed as a percentage. Algebraically, this represents, d/p ( --- _ 1 ) x 100 where pm and pd are pr ies paid by Madagascar dnd industrial countries, respectively, and a_ is the quantity of Madagas lra's imprt.t The computations are based on products for which five-digit unit values were coiputed. See Table 2. b Observatton falling in the top decrle of Table 4, i.e., items with a Madagascar price relative exceeding 150 percent Note: No exutreme price relatives fel Iin the following two-digi t SITC groups - SITC 53 (Dyes and Tanning Products); SITC 73 (Metalworlrng tach nery); SITC 81 (PLumbing and Lighting Fixtures); SITC 85 (footwear); SITC 88 (Photographic Equpinient and Suppt les) 16 (SITC 520). Preshipment inspection is required for almost all Madagascar's imports of products classified in SITC 5 and 6 (some specialty chemicals are excluded--see footnote 7) so the data do not suggest that extreme prices occur primarily where PSI is absent. A second possible cause of industry differences in relative import prices is collaborative false invoicing by foreign exporters and Madagascar importers. This potential explanation recognizes that the incentive to over- or underinvoice depends on the relative height of the foreign exchange black market premium and the tariff rate. If the black market premium is relatively high this encourages overinvoicing to facilitate capital flight, while a relatively high tariff encourages underinvoicing to minimize import duties.17 Expressed algebraically, if t1 is the nominal tariff, p is the Dlack market premium (measured as a percentage above the official exchange rate), V, is the true value of imports, and Vf is the falsified invoice price, then the importer's net gain (or loss) on product i (Ni) will equal, (5) N1 = t1 (Vt - Vf) - p(Vt - Vf) or, 17For example, if the tariff is 70 percent and the value were underinvoiced by 25 percent, the importer actually pays a 52.5 percent duty. This assumes, however, that the importer can obtain foreigr exchange to finance that part of the import bill which is underinvoiced. If exchange ccntrols exist, the extra foreign exchange must be purchased on .he black market at a premium over the official rate. In this situation underinvoicing is profitable if the tariff exceeds the black market premium. It follows that goods with very high tariffs- -say 100 percenc or more--are the most likely to be vehicles for tax evasion by underinvoicing. It should also be noted that the statistical tests presented in this paper, which are based on French export unit values, assume that exporters and importers collaborate on the false invoicing. It may be that the misirvoicing is done solely by importers (if it occurs) and the French export data accurately reflect relative prices charged. 17 (6) N1 = (ti - p) LVt - Vf Equation (6) shows that if p > t, importers have an incentive to overinvoice, Vf > Vt. If p < t, the incentive operates in the reverse direction. According to Pryor (1988, p. 37) Madagascar's black market premium ranged from 50 to 70 percent during 1982-84. Appendix Table 2 gives tariff rates for different imports. Since as equation (6) shows, the relative level of tariffs and the black market premium determine the direction of incentives for false invoicing, data on the level of Madagascar's tariffs and related import charges were drawn from an UNCTAD (1987) report. This source gives nominal import duty averages for many SITC (Revision 2) products down to the five-digit level. Using these statistics, items were ranked by decreasing tariffs and two groups selected. The first was composed of high tariff items (import duties for this group ranged from 40 to 131 percent--see Appendix Table 2) while the second consists of products with relatively low (15 percent or under) tariffs and special import charges.18 Next, Madagascar's average relative import price was computed for each group before and after PSI was adopted. The results are reported in Table 6. 18The complete list of products included in these two groups along with their corresponding SITC (Rev. 2) codes and nominal tariffs is presented in Appendix Table 2. Over the period covered by these tests Madagascar's black market premium and tariffs changed in ways that would be offsetting. In 1987 and 1988 .mport tariffs were simplified and reduced and quantitative restrictions eliminated. According to estimates by Pryor (1988, p. 37) government policies restored economic growtn and cut the black market exchange rate by about one- third from its 1978-80 level. 18 Table 6: Relative Import Prices for High and Low Madagascar Tariff Products: Based on Statistics Reported by Industrial Countries (Price Relatives Using Industrial Country Comparisons, Low Tariff Products High Tariff Products Exporter 1979-82 1984-88 1979-82 1984-88 France 71.3 49.0 66.4 25.9 Germany, Italy and Japan 199.6 75.1 185.1 54.8 All Abov~e 125.0 70.2 98.9 44.6 Note: For all of the above comparisons except France, over the 1979-82 period the low tariff product average price relative is significantly higher (95 percent confidence level) than the high tariff product average. Also, the 1984-88 price relatives for both the high and low tariff products are significantly lower than the 1979-82 figures. The reader should note these results were achieved with a smaller and less repre- sentative sample than the findings reported in Table 3. The data in the table support the proposition that collaborative over- invoicing contributes to product differences in price relatives. 19 In both 1979- 82 and 1984-88 low tariff products had significantly higher prices (the 1979-82 pre-PSI difference for French exports was not significant) which is the pattern expected under collaborative false invoicing. Moreover, the spread between the high and low tariff products' price relatives (about 26-percentage points) is 19It should be noted that the results are also consisten with other possible explanations. For example, high tariff items are concentrated in labor- intensive sectors where Madagascar has, or could develop, a productive capacity. As such, "potential" competition from domestic producers may moderate foreigners, export prices, as would be the case if "limit" pricing were being practiced (see Yeats, 1976 for a discussion of limit pricing models). Also, the (simpler) labor-intensive, high tariff products may be more "familiar" to customs agents so the potential for inflating their prices is reduced. 19 almost identical in 1979-82 and 1984-88--an observation that suggests PSI did little to diminish the relative importance of false invoicing. IV. Summary and Conclusions To counter the adverse effects of pricing and other trade practices a growing number of developing countries engage preshipment inspection firms to verify quality and quantity standards of traded goods, and to determine if prices are within acceptable norms. The fact that PSI is relatively expensive--costs appear to average about one percent of the value of goods inspected--heightens the need for objective evaluations of PSI. The present study evaluated one objective of preshipment inspection by analyzing Madagascar's relative import prices before and after PSI was introduced. The following conclusions result: Comparisons with average prices charged industrial and developing countries indicate Madagascar paid a premium for most imported goods lefore and after PSI was adopted. Moreover, the data show that preshipment inspection failed to bring Madagascar's prices closer to the average for other importers. Madagascar's inflated import prices under PSI involved major associated revenue losses. If Madagascar paid the same averace prices as other countries after PSI was adopted the savings for chemicals (SITC 5) and basic manufactures (SITC 6) imported from France alone would be on the order of US548 to 52 million, with an associated savings of US$3 to 4 million for iron and steel products (see Appendix 1). If the 30 to 40 percent premium Madagascar paid 20 over 1984-88 is applied to all goods this implies annual losses of US5125 to 150 million. The most extreme overpayments are clustered in chemicals and basic manufactures--areas where preshipment inspection is generally required. Furthermore, there is some (tentative) evidence that collaborative false invoicing exists as import price relatives for high tariff items are significantly below those for low tariff products. However, there are several other alternative hypotheses that could account for this pattern. This study's findings raise several issues that require clarification. First, there is a need for specifics on the price variation that will pass PSI inspection. Do Madagascar's 30 to 50 percent above-average relative prices fall within the acceptable range for the preshipment inspectors? Would results improve if a conscious effort were made to tighten the range? A related question is whether or not the pricing practices in sectors like chemicals, where the extreme (adverse) price relatives are concentrated, are sufficiently complex and secretive that PSI is likely to be relatively ineffective u,nder existing conditions. Several issues of importance were not addressed in this study that warrant further research. First, preshipment inspection is often adopted to address nonprice problems. These include shipment of defective goods or goods that fail to meet contractual standards and quantities. In cases, such as Indonesia, preshipment inspection was used to combat graft, corruption and inefficiency in customs services. How effective PSI is in dealing with these key problems will require "on-site" evaluations of PSI operations. Such evaluations should also attempt to determine if there are unintended effects of the inspection program. 21 Is trado being diverted to suppliers in countries when PSI is not required? Are large consignments being broken up and shipped in smaller units to avoid inspection (which may not be required on imports under a certain value)? Has PSI involved costly new administrative procedures, or has it improved customs procedures and speeded goods chrough import controls? Are there reasonable alternatives to PSI as it is now conducted? These are the types of issues that should be addressed in further research. A final point is that consideration should be given to the nature of the service that would best serve the needs of developing countries--is it PSI or assistance with general Procurement problems. Specifically, preshipment inspection now focuses on ensuring that the contracting party pays a "reasonable" price for goods from a civen country (or receives an adequate price for exports), but does not attempt to identify low(er) cost suppliers. Further assessments of PSI might specifically address this issue by comparing prices actually paid by the contracting country with those charged by alternative suppliers. The data sources and empirical procedures employed in this study could be easily adapted for an evaluation of this question. 22 References Avramovic, Dragaslov. 1978. "Common Fund, Why and What Kind." Journal of World Trade Law. 12 (October). Bhagwati, Jagdish. 1967. "Fiscal Policies, the Faking of Foreign Trade Declarations, and the Balance of Payments." Oxford Bulletin of Economics ind Statistics. (February). Dornbusch, Rudiger. 1987. Over and Underinvoicina: Motives. Evidence and Macroeconomic Implications. Hearings Before the United States International Trade Commission, Investigation No. 332-242. Washington: USITC, monograph. (February). Edwards, Corwin. 1972. "Barriers to International Competition: Interfirm Competitive Behavior," in R. Hawkins and Ingo Wolter, eds. The United States and International Markets. Lexington, Mass.: Heath. Ely, J.E. 1961. "Variations Between U.S. and Its Trading Partner Import and Export Statistics." The American Statistician. (April). Erzan, Refik et. al. 1989. "The Profile of Protection in Developing Countries." UNCTAD Review. 1, No. 1. Helleiner, G. 1978. World Market Imperfections and Developing Countries. Washington: Overseas Development Council. Hufbauer, G.C. and J.P. O'Neill. 1972. "Unit Values of U.S. Machinery Exports.. Journal of International Economics. vol. 2, pp. 265-276. Laird, Samuel and Alexander Yeats. 1990. Ouantitative Methods for Trade Barrier Analysis. London:Macmillan Press. McAllister, Harry. 1961. "Statistical Factors Affecting the Stability of the Wholesale and Consumer Price Indexes." In U.S. Congress, Joint Economic Committee Government Price Statistics Hearing. Washington: U.S. Government Printing Office. Morgetistern, Oskar. 1963. On the Accuracy of Economic Observations. Princeton: Princeton University Press. Mowbray, Patricia de. 1988. "Foreign Exchange Conservation and Tax Efficiency in Developing Countries: The Economic Case for Pre-Shipment Inspection." The Business Economist 20, no. 1 (Winter) 34-48. Pryor, Frederic. 1988. "Income Distribution and Economic Development in Madagascar: Some Historica: Statistics." World Bank Discussion Paper Number 37. Washington: World Bank. Sheikh, Munir. 1974. "Underinvoicing of Imports in Pakistan." Oxford Bulletin of Economics and Statistics. (November). 23 Stigler, George and James Kindahl. 1970. The Behavior of Industrial Prices. New York: Columbia University Press. Simkin, C.G.F. 1970. "Indonesia's Unrecorded Trade." Bulletin of Indonesian Economic Studies. (March). Societe General de Surveillance. 1988. Examples of Results Obtained by Preshioment Inebection and Price Comparison. Geneva:SGS (Septembor). Societe General de Surveillance. 1989. The Validity of SGS Trade Monitorino Programme. Geneva:SGS (March). UNCTAD. 1987. Handbook of Trade Control Measures of Developino Countries. (UNCTAD/DDM/Misc.2) Geneva:UNCTAD. United States International Trade Commission. 1987. Preshioment Inspection Programs and Their Effects on U.S. Commerce. Washington:USITC Publication No. 2003 (August). Walter, Ingo. 1987. "The Mechanism of Capital Flight." In Donald Lessard and John Williamson, eds. Ca2ital Flight and Third World Debt. Washington: Institute for International Economics. World Bank. 1989. Trends in Develovina Economies 1989. (Washington: World Bank). "Madagascar." pp. 254-259. Yeats, Alexander. 1976. "A Framework for Analyzing Potential Competition in Financial Mergers and Acquisitions." Journal of Money Credit and BankLnc. (August). Yeats, Alexander. 1978. "On the Accuracy of Partner Country T:ade Statistics." Oxford Bulletin of Economics and Statistics. (November). Yeats, Alexander. 1990a. "On the Accuracy of Economic Observations: Do Sub- Saharan Trade Statistics Mean Anything?" World Bank Economic Review 4, No. 2 (May) 135-156. Yeats, Alexander. 1990b. "Do African Countries Pay More for Imports?.. .Yes." World Bank Economic Review 4, No. 1 (January) 1-20. 24 Appendix 1 Analysis of Madagascar's Relative ImPort Prices and Expenditure Effects for Five-Digit SITC Iron and Steel Products 1979-82 and 1984-88 25 While the precedir.g analysis employed unit values for all types of goods exported to Madagascar--some of which may be subject to product-mix changes-- there are several product categories wherL. this factor's influence is thlought to be small. Specifically, studies by Stigler and Kindahl (1979), McAllister (1961) and others used iron and steel unit values to assess the accuracy of price quotations employed by the United States Bureau of Labor Statistics for the U.S. Wholesale Price Index, while Yeats (1978) (1990b) employed similar information to identify factors producing differences in international transaction prices. As such, it appears useful to determine if Madagascar's relative prices for these homogenous products follow the same pattern as other goods' prices before and after the adoption of preshipment inspection. Appendix Table 1 provides information on each five-digit SITC (Rev. 2) iron and steel product imported by Madagascar for which 1979-88 quantity and value data were available. The table shows the value of Madaga^car's imports of each item from France in the pre- and post-PSI periods as well as the average unit value for these shipments.' For comparison, unit values for French exports to other developing and industrial countries are also shown. Finally, a summary measure of the expenditure effects of the differences in relative prices was computed. This measure shows the expenditure gain or loss on imports if Madagascar paid the same average prices as other countries. Stated algebraically, lAn attempt was made to compile similar data for other suppliers, i.e., Germany, Italy and the United Kingdom. Two problems were encountered with the major difficulty being that Madagascar primarily imports ferrous metals from France. Also, the other European countries generally did not provide export data at the same level of detail (i.e., for five-digit iron and steel products) and some quantity data required for computation of unit values were missing. Appendix Table 1: Comparative Inport Prices and Expenditure Effects for five-Digit Iron and Steel Products Imported from Franre French unit values for Implied Madagascar gains or French exports to Madagascar exports to other countries tosses based on other's prices (SW0) SITC Descriptions Years Vatue (S000) Unit value (S) Developing Irnustrial Developing Irxustrial - - ALL SAMPLED STEEL ITEMSb 1979-82 28,212 513.62 434.73 377.84 -4,855 -6,969 1984-88 19,131 530.61 406.01 446.71 -4.188 -2,797 673.26 Iron and Steel Hot-Rolted Bars 1979-82 9,195 418.77 351.94 358.21 -1,467 -1,330 1984-88 3,238 359.34 282.81 313.63 -690 -416 673.27 Iron and Steel Forged Bars 1979-82 244 983.87 1,040.54 767.63 14 -54 1984-88 229 978.63 667.92 512.73 -73 lo9 673.31 Iron and Steel Small U.I.H. Sectiens 1979-82 138 572.61 650.88 535.25 19 9 1984-88 179 471.05 473.84 483.89 1 5 673.32 Iron and Steel Large U.I.H. Sections 1979-82 1,436 460.11 357.12 366.58 -321 -292 1984-88 418 277.37 252.94 253.63 -37 -36 673.33 Iron and Steel Hot-Rolled Profiles 1979-82 2,397 470.46 440.40 401.49 153 351 1984-88 851 329.21 411.39 352.18 212 59 ,J 673.36 Iron and Steel Simpte Sheet Piling 1979-82 *25 '51.78 426.18 457.28 -74 56 1984-88 41' 620.27 394.70 479.19 -149 -93 674 41 Heavy Ptates of Iron or Simple Steel 1979-82 873 510.23 392.35 362.69 -202 252 1984-88 1,294 460.01 362.42 295.05 -275 464 674.51 Medium Plates of Iron or Siople Steel 1979-82 664 497.01 345.67 448.18 -202 65 1984-88 746 445.90 383.89 413.72 -104 54 674.61 Rolled Thin Plate of Iron or Steel 1979 82 1,271 492.44 362.34 412.08 -336 207 1984-88 2,891 456.88 335.39 424.82 7,71 203 674.91 Other Iron and Steel Plates 1979-82 3,894 824.30 583.95 504.89 1,135 1,509 1984-88 8,737 654.80 481.33 545.95 2,315 1,452 676.01 Iron and Steel Railway Rails 1979 82 7,775 476.35 415.00 302.09 1,001 -2,844 1984-88 131 423.95 465.69 313.11 13 -34 a The calculations are based on the actual expenditure by Madagascar minus the expenditure that would have been required it Madagascar paid the same prices as other inporters. Algebraically, the inplied gain or loss (Ege) is: Ege = Pf% - PM% here Pf and pM are prices paid by foreign and Madagascar importers and c is the quantity of Madagascar imports. b The aggregate unit values have been com,uted using Madagascar's trade weights for the appropriate time period. Eg - ( Ptn - p.nm) where P# and Pm are the prices paid Ly ct!.er (t:reig. traziers an Madagascar, respectively, !_r Frenrz exp,;rts, anJ u,l is tne q,arti,ty _. t-e f.ve-digit good imuported by Madagascar. A regat ve .a _e re resents :_-netary ._sses caused by hiozher Madagascar prices while a positive aue .ndicates a ga r. from relatively lower import prices. The comparisons with other developing countries provide no evidence that Madagascar's relative import prices improved after PSI was adopted; a finding that matches the conclusions based on all imports (see Tables 2 through 4). In fact, Madagascar's relative prices rose to 23.5 percent above the average for other developing countries after PSI was required--up about 8-percentage points from average 1979-82 prices. The relative price differences imply expenditure losses for Madagascar of USS4.2 million--down from the 1979-82 losses of US$4.9 million on a considerably larger import base. Relative to the industrial countries, Madagascar's import prices were still 16 percent higher in the 1984- 88 period although the associated excenditure losses declLned by over 50 percent (to about US52.8 million). To a large degree thIs was due to a compositional change in. Madagascar's imports--particularly the reductisn in iron and steel railway rails (SITC 676.01) where Madagascar was at a ma,or competitive price disadvantage in 1979-82.2 All in all, the evidence from Appendix Table 1 is 2For example, over 1979-82 Madaciascar imported 16,322 tons of steel rails at an average price of USS476.35 per ton--as opposed to a price of USS302.09 for industrial countries. Had Madagascar imported this same quantity under the 1984- 88 relative prices the implied expenditure loss would have been US$1.8 million rather than the US$34 thousand reported in the table. Appendix Table 1 shows that some 1984-88 unit values were lower than they were in 1979-82. This is due to the UN practice of converting all trade data to dollars, and the appreciation of the dollar against the French franc in the mid-1980s. In other words, French- franc prices rose from 1979-82 to 1984-88 but they appear lower in dollar terms. 28 cons9itent with previous find.Lngs tnat PS: did r.t sg.nf".:-art.y .rprove Madagascar's relative irrpcrt prices. 29 Appendix 2 High and Low Tariff Products for Tests of False Invoicing by Industrial Country Exporters and Countries Covered by M3daaascar's PSI Proaram 3 :- Appendix Tab.e 2: H!Y , . j- T.+< _e, A--j W- , * .e .; T3- + ad S- aP!'> -r narates) AIe'a;e -a-' i Ra"ge Ave. TOtal SITC(Rev.2) Cescriptj,n Tarift M:r'''wjii Max :ixmur Cnarges I. HIGH TARIFF PROC..CTS 01 Meat and preparations 6.1 5.0 10.0 106.1 03 Fish, crustaceans, etc. 3.6 0.0 15.0 101.7 05 Vegetables and fr't 9.1 0.3 20.0 101.9 06 Sugar and preparat:cns 11.9 0.C 15.0 63.1 07 Coffee, tea, cocoa, spices 14.5 1S.0 20.0 130.7 09 Edible products, nes 8.3 0.0 20.0 84.0 24 Cork and wood 0.9 0.0 5.0 70.7 41 Animal fat and oil 4.4 0.0 5.0 43.3 53 Dyeing and tanning material 3.4 0.0 10.0 40.4 55 EssentiaL Oits 6.5 0.0 15.0 60.3 57 ExpLosives and pyrotechnics 7.3 5.0 15.0 48.0 61 Leather and dressed skins 6.2 5.0 10.0 45.2 625.2 New tires 7.5 5.0 10.0 40.5 63 Cork and wood manufactures 4.3 0.0 5.0 54.9 65 Textile yarn and fabrics 11.0 0.0 40.0 56.6 6.6(ess 661.2) Norvnetalic mineraL manufactures 7.2 0.0 15.0 40.9 716.23 Generators with piston engines 10.0 10.0 10.0 41.0 741.5 Air conditioning machines 10.0 10.0 10.0 46.0 743.6 Gas, liquid filters, etc. lS0 10.0 50.0 40.0 749.1 Ball and roLter bearings 10.0 10.0 10.0 41.0 749.3 Transmission shafts 10.0 10.0 10.0 41.0 75 Office machines 10.5 9.0 15.0 47.6 76 Telecormunications eouioment 5.0 5.0 5.0 63.2 776.4 Electronic microcircuits 5.u 5.0 5.0 65.0 781.0 Passenger mctor ve'Cles 1¶7.5 10S.0 20.0 94.3 782.1 Lorries, trucKs, etc. 12.0 10.0 20.0 56.8 784.9 Parts of motor vehicles 10.0 10.0 10.0 41.0 81 Sanity fixtures 7.0 5.0 10.0 43.4 82 Furniture and parts 8.6 5.0 10.0 94.3 83 Travel goods 10.0 10.0 10.0 85.0 84 Apparet 19.2 0.0 25.0 71.8 85 Footwear 10.0 10.0 10.0 59.1 88 Optical goods and watches 11.3 5.0 20.0 56.7 89(less 892.11) Misc. manufactured goods 6.1 C." 15.0 55.7 II. LOW TARIFF PROCUCTS 022.42/3 Dry and Dowdered milk 2.5 0.0 5.0 14.5 041.1/2 Duruim and other wheat 5.0 5.0 5.0 13.0 042.21 Milled rice 0.0 0.0 0.0 2.0 046.01 Ftour of wheat 5.0 5.0 5.0 13.0 08 Animal feeds 5.0 0.0 10.0 11.7 12 Tobacco and manufactures 5.4 0.0 15.0 5.9 23 Crude rubber U.0 0.0 0.0 15.0 32 Cual and coke 0.0 0.0 0.0 9.1 33 Petroleuti products 0.2 0.0 5.0 14.7 34 Gas, natural and manufactured 0.0 0.0 0.0 7.8 56 Manufactured fer:ilizers 0.0 0.C 0.0 0.0 661.2 Cement 5.0 5.0 5.0 10.0 672.71 Iron and steel coils 0.0 0.0 0.0 12.0 723.9 Parts of construction machines 5.0 5.0 5.0 15.0 744.28 Other handling machines 5.0 5.0 5.0 15.0 782.2 Special purpose vehicles 10.0 10.0 10.0 13.0 79 Other transport equipmrent 3.5 0.0 10.0 16.3 892.11 Printed books 1.7 0.0 5.0 11.0 Source: UNCTAD, Handbook of Trade ControL Measures of DeveloDng Countries (UNCTAD/DDM/Misc.2) (Geneva:UNCTAD 1987), pp. 190-191. The UNCTAD source provides ta;riff data for all SITC Rev. 2 two-digit headings as well as for 100 most important five-digit products imported by developing countries. 3 1 Appendix Table 3: Countries in which Pres;-spme,- :ts O-curs for Expcrts Eurooe Asia Africa The Americas Austria Bangladesh A'geriaa Argentina Belgium Burmaa Egypt3 Bolivia Bulgariaa Hong Kong Ghana Brazil Czechoslovakiaa India Ivory Coast Canada Denmark Indonesia Kenya Chile F.R. Germany Iran Malawi Colombia Finland Israel Morocco Costa Rica France Japan Mozambiquea Cubaa German D.R.a Kuwait Nigeria Ecuador Greece Lebanon Tanzania El Salvador Holland Malaysia Tunisia Mexico Hungarya Pakistan Zambia Panama Iceland Philippines Zimtabwe Paraguay Italy Saudi Arabia Peru Luxembourg Singapore Puerto Rico Malta South Korea Tr.Ln1dad and Tobago Norway Sri Lanka U.S.A. Poland Thailand bruguay Portugal Venezuela Romaniaa Spain Sweden Switzerlandb Oceania Turkey Australia United Kingdom New Zealand Yugoslaviaa a Countries in which SGS performs the quantlty and quality Inspection, but not the price comparison. b Special Swiss system. PRE Working Pager Series Contact 1B9 A&= naDfr WPS596 The Mexican Sugar Industry: Brent Borrell February 1991 P. Kokila Problems and Prospects 33716 WPS597 Rent Sharing in the Mufti-Fibre Refik Erzan February 1991 G. Ilogon Arrangement: Theory and Evidence Kala Krishna 33732 from U.S. Apparel Imports from Ling Hui Tan Hong Kong WPS598 Africa Region Population Projections: Patience W. Stephens February 1991 0. Nadora 1990-91 Edition Eduard Bos 31091 My T. Vu Rodolfo A. Bulatao WPS599 Asia Region Population Projetions: Eduard Bos February 1991 0. Nadora 1990-91 Edition Patience W. Stephens 31091 My T. Vu Rodolfo A. Bulatao WPS600 Latin America and the Caribbean My T. Vu February 1991 0. Nadora Region Population Projections: Eduard Bos 31091 1990-91 Edition Patience W. Stephens Rodolfo A. Bulatao WPS601 Europe, Middle East, and North Eduard Bos February 1991 0. Nadora Africa Region Population Projections: Patience W. Stephens 31091 1990-91 Edition MyT.Vu Rodolfo A. Bulatao WPS602 Firm Output Adjustment to Trade Mark A. Dutz February 1991 S. Fallon Liberalization: Theory with 38009 Application to the Moroccan Experience WPS603 The Role of Officially Supported Asli Demirguc-Kunt February 1991 G. llogon Export Credits in Sub-Saharan Refik Erzan 33732 Africa's External Financing WPS604 Foreign Trade and Its Relation to Faezeh Foroutan February 1991 S. Fallon Competition and Productivity in 37942 Turkish Industry WPS605 Overview of Gontractual Savings Dimitri Vittas March 1991 W. Pitayato- Savings Institutions Michael Skuiiy nakarn 37666 WPS606 Adjustment Policies and Investment Luis Serven March 1991 E. Khine Performance in Developing Andres Solimano 39361 Countries: Theory, Country Experiences, and Policy Implications PRE Working Paper Series Contact ilii Alo for pager WPS607 Abolishing Green Rates: The Effects Donald F. Larson March 1991 D. Gustafson on Cereals, Sugar, and Oilseeds Simon Glance 33714 in West Germany Brent Borrell Merlinda Ingco Jonathan Coleman WPS608 Cross-Country Studies of Growth Ross Levine March 1991 CECMG and Policy: Methodological, David Reneit 39175 Conceptual, and Statistical Problems WPS609 A Sensitivity Analysis of Cross- Ross Levine March 1991 CECMG Country Growth Regressions David Renelt 39175 WPS610 Can Preshipment Inspection Offset Alexander J. Yeats March 1991 J. Jacobson Noncompetitive Pricing of Developing 33710 Countries' Imports? The Evidence from Madagascar