EXPANDING THE INFORMATION TECHNOLOGY AGREEMENT (ITA)

Transcription

1 DG Trade EXPANDING THE INFORMATION TECHNOLOGY AGREEMENT (ITA) ECONOMIC AND TRADE IMPACTS APPENDIX TO FINAL REPORT OCTOBER 2010

2 APPENDIX 1 BACKGROUND OF THE ITA AGREEMENT The Ministerial Declaration on Trade in Information Technology Products (ITA) was concluded by 29 participants at the Singapore Ministerial Conference in December The number of participants has grown to 70, representing about 97 per cent of world trade in information technology (IT) products. 1 The ITA provides for participants to completely eliminate duties on IT products covered by the Agreement. Developing country participants have been granted extended periods for some products. The ITA is solely a tariff cutting mechanism. While the Declaration provides for the review of non-tariff barriers (NTBs), there are no binding commitments concerning NTBs. There are three basic principles that one must abide by to become an ITA participant: 1) all products listed in the Declaration must be covered, 2) all tariffs must be reduced to a zero tariff level, and 3) all other duties and charges must be bound at zero. There are no exceptions to product coverage, however for sensitive items, it is possible to have an extended implementation period. The commitments undertaken under the ITA in the WTO are on an MFN basis, and therefore benefits accrue to all other WTO Members (critical mass agreement). Table A1.1 List of ITA countries ITA Countries Albania Czech Rep. Guatamala Jordan Morocco Romania Ukraine Australia Denmark Honduras Korea Netherlands Saudi Arabia United Arab Emirates Austria Dominican Rep. Hong Kong Kyrgyz Rep. New Zealand Singapore United Kingdom Bahrain Egypt Hungary Latvia Nicaragua Slovak Rep. United States Belgium El Salvador Iceland Lithuania Norway Slovenia Vietnam Bulgaria Estonia India Luxembourg Oman Spain Canada Finland Indonesia Macoa Panama Sweden China France Ireland Malaysia Peru Switzerland Costa Rica Georgia Israel Malta Philippines Taiwan Croatia Germany Italy Mauritius Poland Thailand Cyprus Greece Japan Moldova Portugal Turkey 1 For more information on the ITA see 2

3 APPENDIX 2 DESCRIPTION OF THE PARTIAL EQUILIBRIUM MODEL The overall purpose of the partial equilibrium analysis is to assess the welfare impacts of changing the bilateral tariff structure on ITA products. The analysis covers 34 countries and/or country groupings. Here, we include the largest importers and exporters of ITA products as well as four groups of countries A2.1 THE THEORETICAL FOUNDATION OF THE PARTIAL EQUILIBRIUM MODEL The welfare impact of eliminating tariffs on international trade depend on the supply, demand and substitution elasticities applied in the model welfare, cf. Box A2.1. Box A2.1 The impact on welfare from imposing a tariff on ITA products When the EU eliminates tariffs on imports from other ITA countries, for example, EU producers would obtain a producer surplus equal to A+B and EU consumers would obtain a consumers surplus equal to C if there was no import from other ITA countries. By importing ITA products from the other ITA countries, the EU consumer ends up paying a lower price and consuming larger amounts of ITA products than would have been the case if only EU products were available. This means that the consumer surplus is now C+B+D. The lower demand for EU products reduces the producer surplus by B. While there is a transfer from producers to consumers worth of B, we also note that there is a gain in welfare equal to D. By imposing an import duty (where t is the tariff rate) on ITA products originating from other ITA countries, we see that the welfare gain falls from D to D. While some of the lost welfare is recaptured as tariff income (equal to D ) we also see that the EU Community will experience a net welfare loss in welfare equal to E +E. The figure underneath has been drawn under the assumption that the EU and the ITA supply elasticities are the same, which is the case in our empirical model. If the domestically produced and the imported products are similar then the tariff will eliminate the full difference between the EU and the ITA price. However, if there are quality differences or other factors that make the two products less than perfect substitutes (e.g. non-tariff barriers) then the full welfare potential from eliminating tariffs will not be fully exploited, and the ITA supply curve will remain below the EU supply curve. Price Price S EU S EU C C P EU P ITA A B D S ITA P EU B D P ITA B D A E E 1+t S ITA D EU D EU Quantity Quantity The welfare impacts depend on the elasticities applied in the model: The larger the supply elasticity (the flatter the supply curve) the larger the welfare gain from eliminating EU tariffs on imports from other ITA countries. This is so because the lower price has reduced supply and thereby EU consumption. 3

4 The larger the EU demand elasticity (the flatter the EU demand curve), the larger the change in welfare of eliminating tariffs on imports from non-eu ITA countries. This is so because the even a small reduction in tariffs will lead to a large increase in demand and EU consumption. The larger the substitution ution elasticity (the smaller the vertical distance between EU and ITA countries) the larger the welfare gain from eliminating EU tariffs on imports from other ITA countries. This is so because EU demand can more easily shift towards imported products and take advantage of the cheaper products made available by the tariff reduction. A2.2 DATA REQUIREMENTS IN THE PARTIAL EQUILIBRIUM MODEL The partial equilibrium analysis requires a vast amount of data. The bilateral matrix of trade flows in value terms Trade flows are expressed c.i.f and cover both the new ITA products (for the product extension scenario) as well as the present ITA products (for the country extension scenario). All trade data is based on the UN Comtrade Database and is retrieved through the WITS Interface. 2 The analysis excludes trade for which there is not information on bilateral tariffs. This means that the product extension scenario covers only 93 percent of the total trade in the extended list of ITA products. Likewise, the country extension scenario covers only 87 percent of total trade in the present list of ITA products. For the country groups we use the total trade for all the countries in the group. EU import from China, for example, is the sum of import from China for all the Member States that do not enter the analysis individually. For the product groups we use total trade for each 2-digit heading. The value of domestic production for own use This data enters as the diagonal in the bilateral trade matrix. We use information from input-output tables for selected EU countries to get an estimate of how production is divided between own consumption and export. We find that EU countries typically consume 40 percent of their own production and export 60 percent abroad, cf. Table A See Wits.worldbank.org. 4

5 Table A2.1 Share of own consumption in own production in selected EU countries Germany France Netherlands Sweden United Kingdom Average Office machinery and computers 33% 47% 7% 54% 40% 36% Electrical machinery and apparatus n.e.c. 61% 58% 45% 51% 63% 55% Radio, television and communication equipment and apparatus 45% 53% 15% 30% 53% 39% Medical, precision and optical instruments, watches and clocks 28% 58% 19% 49% 58% 42% Average 42% 54% 22% 46% 53% 43% Note: The figures show the share of own production that is used domestically. The rest of the production is exported to other countries. Source: Own calculations based on input-output tables downloadable from For the country groups the diagonal also includes also intra-group trade. For the EU, for example, the diagonal therefore includes (i) consumption of own production for all the EU Member States that do not enter the analysis individually and (ii) trade between the Member States that do not enter the analysis individually. In this way, we ensure that our model covers all consumption of ITA products. The matrix of bilateral tariffs Tariff data is downloaded through the WITS interface from the TRAINS database. The tariffs are measured as effective import tariff rates by HS 4-digit products, where the 4-digit level is a weighted average calculated by the TRAINS database. For the product groups (84, 85 and 90) we calculate a weighted average where the weights applied are the import share of the individual product included in the product group. Elasticities of supply, import demand, and substitution The model requires data on supply, demand and substitution elasticities. The supply elasticity measures how much supply in a given country responds to changes in the price of imported goods. Since the production of ITA products does not depend on scarce production factors such as natural resources we expect supply to be infinite (equal to 100 in the model). This means that an increase in the demand for ITA product will always be matched by an expansion of production. The implication is that increased export does not take place at the expense of domestic consumption. The elasticity of import demand measures how much the demand for imports decreases when the price of imported goods increases. We use the GTAP elasticity of substitution for manufactures and electronics (average across the countries in the sample) equal to The demand elasticities can be found at 5

6 The elasticity of substitution measures the degree of substitution between a domestically produced and an imported product, i.e. by how many percentages demand for domestically produced goods increases as the price of imports increases by 1 percent. The GTAP elasticity of substitution for electronic equipment equal to 4.4 (available at reflects the short run substitution between domestically produced and imported ITA products. The long run substitution elasticity is expected to be higher since imported goods have the opportunity to adjust to domestic demand. As stated in USICT (2002) 3 Long-run estimates are up to five times as large as short-run estimates, and on average twice as large as the short-run estimates. This is important since long-run estimates are more appropriate for most trade policy analysis than short-run estimates To be conservative we use a substitution elasticity equal to 6 which is larger than the shortrun GTAP elasticity equal to 4.4 less than twice as large (and far below a long-run elasticity five times larger than the short-run elasticity). There are good reasons for being conservative because, as we will see below, the results are rather sensitive to changes in the substitution elasticity. The elasticity of substitution is assumed to be constant across all country pairs. Table A2.2 Table A2.4 show the results from eliminating tariffs on an extended number of ITA products under HS heading 85 when we apply different elasticities. 4 We find that the net results range between 44.3 million and million. The simulation result when we use our chosen elasticities gives a net welfare gain worth 101 million lies well in the middle of these results. Table A2.2 Sensitivity of the net welfare gain with respect to the demand elasticity Scenario 1 Scenario 2 Scenario 3 Demand elasticity Substitution elasticity Supply elasticity Welfare impact Table A2.3 Sensitivity of the results with respect to the substitution elasticity Scenario 1 Scenario 2 Scenario 3 Demand elasticity Substitution elasticity Supply elasticity Welfare impact See USICT (2002), A Discussion on Armington Trade Substitution Elasticities, downloadable at 4 This product group has been chosen since it constitute the bulk of EU export of ITA products. 6

7 Table A2.4 Sensitivity of the results with respect to the supply elasticity Scenario 1 Scenario 2 Scenario 3 Demand elasticity Substitution elasticity Supply elasticity Welfare impact The employment shares of value added To carry out the partial equilibrium model, we need data on the employment share of value added for the ITA industry and the downstream industries (i.e. the users of ITA products). For the OECD countries, data is available on the ISIC Rev. 3 level from the OECD STAN database for structural analysis. For the remaining countries we use the OECD average. There is no official translation between the HS codes used for the trade and tariffs matrices and the ISIC Rev. 3 industries. For the HS32 and HS37products we use ISIC Rev. 3 code 24 (chemicals and chemical products), HS39 products are proxied by ISIC Rev. 3 code 25 (rubber and plastic products), for the HS74 products we use ISIC Rev. 3 code 27 (basic metals), for the HS84 and HS85 products we use ISIC Rev. 3 code 31 (electrical machinery and apparatus), and for the HS90 products we use ISIC Rev. 3 code 33 (medical, precision and optical instruments).for the downstream we use data on the whole manufacturing sector. There is little variation in the employment share of value added across industries, cf. Figure A2.1. Also, there is little difference between the protected ITA industries and the downstream manufacturing industry. Figure A2.1 Value added across industries Employment share of value added in the protected and downstream industries Note: The employment share of value added is calculated as total labour costs divided by value added. Source: OECD STAN database on structural analysis. 7

8 Our grouping of countries Our partial equilibrium analysis consists of 30 individual countries and four country groupings as listed in Table A2.5 below. The groups are formed in order to ensure that the main ITA trading partners enter on an individual basis. Table A2.5 Groups of countries in the partial equilibrium model Individual countries Other EU countries Other ITA Countries Other WTO non-ita member Antigua and Barbuda Austria Bulgaria Albania Mongolia Rest of the world Brunei Darussalam Belgium Cyprus Australia Argentina Mozambique Côte d'ivoire Brazil Czech Republic Costa Rica Armenia Namibia Egypt Canada El Salvador Croatia Bahrain Niger FYR of Macedonia Dominican Republic China Estonia Bangladesh Nigeria Macoa Denmark Greece Georgia Barbados Pakistan Saint Lucia Saint Vincent & Finland Latvia Guatemala Belize Panama the Grenadines France Lithuania Honduras Bolivia Paraguay Sri Lanka Hong Kong, China Germany Luxembourg Botswana Peru Tanzania Hungary Malta Iceland Burundi Qatar The Gambia Indonesia Portugal India Cameroon Rwanda Vietnam Ireland Romania Jordan Cape Verde Senegal Israel Slovenia Kyrgyz Republic Chile Solumon Islands Italy Mauritius Colombia South Africa Japan Moldova Cuba Togo Trinidad and Tobago Korea Morocco Dominica Malaysia New Zealand Ecuador Tunisia Mexico Nicaragua Fiji Uganda Netherlands Norway Gabon Ukraine Philippines Oman Ghana Uruguay Poland Saudi Arabia Grenada Venezuela Singapore Taiwan, China Guinea Zambia Slovak Republic United Arab Emirates Guyana Spain Maldives Jamaica Sweden Switzerland Thailand Turkey United Kingdom United States Kenya Kuwait Madagascar Malawi Mali 8

9 A2.3 THE IMPACT OF EXTENDING THE ITA PRODUCT LIST The impact of including Printing, drawing and writing ink on the ITA list can be found in Table A2.6. Table A2.6 Summary table for printing, drawing and writing ink (Product 32) producer surplus consumer surplus tariff revenue Net change in welfare EU Non-EU ITA Non-ITA WTO Rest of the world Total Note: Numbers reflect the impacts of extending the number of products on the ITA list ( million). The impacts of eliminating tariffs on Chemical preparations for photographic uses can be seen in Table A2.7. Table A2.7 Summary table for chemical preparations for photographic uses (Product 37) producer surplus consumer surplus tariff revenue Net change in welfare EU Non-EU ITA Non-ITA WTO Rest of the world Total Note: Numbers reflect the impacts of extending the number of products on the ITA list ( million). The impacts from extending the ITA list to cover Other articles of plastic can be found in Table A2.8. Table A2.8 Summary table for other articles of plastic (Product 39) producer surplus consumer surplus tariff revenue Net change in welfare EU Non-EU ITA Non-ITA WTO Rest of the world Total Note: Numbers reflect the impacts of extending the number of products on the ITA list ( million). 9

10 Impacts of eliminating tariffs on Parts and accessories of refined copper are listed in Table A2.9. Table A2.9 Summary table for parts and accessories of refined copper (Product 74) producer surplus consumer surplus tariff revenue Net change in welfare EU Non-EU ITA Non-ITA WTO Rest of the world Total Note: Numbers reflect the impacts of extending the number of products on the ITA list ( million). The impact on net welfare from including products under Machinery and mechanical appliances on the ITA list can be seen in Table A2.10. Table A2.10 Summary table for machinery and mechanical appliances (Product 84) producer surplus consumer surplus tariff revenue Net change in welfare EU Non-EU ITA Non-ITA WTO Rest of the world Total Note: Numbers reflect the impacts of extending the number of products on the ITA list ( million). 10

11 APPENDIX 3 CONTRY RESULTS FROM THE PARTIAL EQUILIBRIUM MODEL A3.1 THE IMPACT OF EXTENDING THE NUMBER OF ITA COUNTRIES Table A3.1 ITA country expansion, country-specific results producer surplus consumer surplus tariff revenue Net change in welfare Country group Argentina Non-ITA WTO Australie ITA Austria EU Belgium EU Brazil Non-ITA WTO Canada ITA Switzerland ITA Chile ITA China ITA Germany EU Spain EU Finland EU France EU Hong Kong ITA Hungary EU Indonesia ITA India ITA Italy EU Japan ITA Korea ITA Mexico Non-ITA WTO Malaysia ITA Netherlands EU Norway ITA Other EU EU OTHER NON-EU ITA ITA OTHER WTO, NON-ITA Non-ITA WTO Phillippines ITA Rest of world ROW Russia ROW Singapore ITA Thailand ITA United Kingdom EU USA ITA South Africa Non-ITA WTO Note: Numbers reflect the impact of extending the number of ITA countries ( million). 11

12 A3.2 THE IMPACT OF EXTENDING THE NUMBER OF ITA PRODUCTS Table A3.2 Printing, drawing and writing ink (Product 32) Country producer surplus consumer surplus tariff revenue Net change in welfare Country group Australia ITA Austria EU Belgium EU Brazil Non-ITA WTO Canada ITA Switzerland ITA Chile Non-ITA WTO China ITA Germany EU Spain EU Finland EU France EU Hong Kong ITA Hungary EU Indonesia ITA India ITA Israel ITA Italy EU Japan ITA Korea ITA Mexico Non-ITA WTO Malaysia ITA Netherlands EU Norway ITA Other EU EU OTHER NON-EU ITA ITA OTHER WTO, NON-ITA Non-ITA WTO Philippines ITA Rest of the world ROW Russia ROW Singapore ITA Thailand ITA Turkey ITA United Kingdom EU USA ITA Note: Numbers reflect the impacts of extending the number of products on the ITA list ( million). 12

13 Table A3.3 Chemical preparations for photographic uses (Product 37) Country producer surplus consumer surplus tariff revenue Net change in welfare Country group Australia ITA Austria EU Belgium EU Brazil Non-ITA WTO Canada ITA Switzerland ITA Chile Non-ITA WTO China ITA Germany EU Spain EU Finland EU France EU Hong Kong ITA Hungary EU Indonesia ITA India ITA Israel ITA Italy EU Japan ITA Korea ITA Mexico Non-ITA WTO Malaysia ITA Netherlands EU Norway ITA Other EU EU OTHER NON-EU ITA ITA OTHER WTO, NON-ITA Non-ITA WTO Philippines ITA Rest of the world ROW Russia ROW Singapore ITA Thailand ITA Turkey ITA United Kingdom EU USA ITA Note: Numbers reflect the impacts of extending the number of products on the ITA list ( million). 13

14 Table A3.4 Other articles of plastic (Product 39) producer consumer tariff Net change Country Country surplus surplus revenue in welfare group Australia ITA Austria EU Belgium EU Brazil Non-ITA WTO Canada ITA Switzerland ITA Chile Non-ITA WTO China ITA Germany EU Spain EU Finland EU France EU Hong Kong ITA Hungary EU Indonesia ITA India ITA Israel ITA Italy EU Japan ITA Korea ITA Mexico Non-ITA WTO Malaysia ITA Netherlands EU Norway ITA Other EU EU OTHER NON-EU ITA ITA OTHER WTO, NON-ITA Non-ITA WTO Philippines ITA Rest of the world ROW Russia ROW Singapore ITA Thailand ITA Turkey ITA United Kingdom EU USA ITA Note: Numbers reflect the impacts of extending the number of products on the ITA list ( million). 14

15 Table A3.5 Parts and accessories of refined copper (Product 74) Country producer surplus consumer surplus tariff revenue Net change in welfare Country group Australia ITA Austria EU Belgium EU Brazil Non-ITA WTO Canada ITA Switzerland ITA Chile Non-ITA WTO China ITA Germany EU Spain EU Finland EU France EU Hong Kong ITA Hungary EU Indonesia ITA India ITA Israel ITA Italy EU Japan ITA Korea ITA Mexico Non-ITA WTO Malaysia ITA Netherlands EU Norway ITA Other EU EU OTHER NON-EU ITA ITA OTHER WTO, NON-ITA Non-ITA WTO Philippines ITA Rest of the world ROW Russia ROW Singapore ITA Thailand ITA Turkey ITA United Kingdom EU USA ITA Note: Numbers reflect the impacts of extending the number of products on the ITA list ( million). 15

16 Table A3.6 Machinery and mechanical appliances (Product 84) Country producer surplus consumer surplus tariff revenue Net change in welfare Country group Australia ITA Austria EU Belgium EU Brazil Non-ITA WTO Canada ITA Switzerland ITA Chile Non-ITA WTO China ITA Germany EU Spain EU Finland EU France EU Hong Kong ITA Hungary EU Indonesia ITA India ITA Israel ITA Italy EU Japan ITA Korea ITA Mexico Non-ITA WTO Malaysia ITA Netherlands EU Norway ITA Other EU EU OTHER NON-EU ITA ITA OTHER WTO, NON-ITA Non-ITA WTO Philippines ITA Rest of the world ROW Russia ROW Singapore ITA Thailand ITA Turkey ITA United Kingdom EU USA ITA Note: Numbers reflect the impacts of extending the number of products on the ITA list ( million). 16

17 Table A3.7 Electrical machinery and equipment (Product 85) Country producer surplus consumer surplus tariff revenue Net change in welfare Country group Australia ITA Austria EU Belgium EU Brazil Non-ITA WTO Canada ITA Switzerland ITA Chile Non-ITA WTO China ITA Germany EU Spain EU Finland EU France EU Hong Kong ITA Hungary EU Indonesia ITA India ITA Israel ITA Italy EU Japan ITA Korea ITA Mexico Non-ITA WTO Malaysia ITA Netherlands EU Norway ITA Other EU EU OTHER NON-EU ITA ITA OTHER WTO, NON-ITA Non-ITA WTO Philippines ITA Rest of the world ROW Russia ROW Singapore ITA Thailand ITA Turkey ITA United Kingdom EU USA ITA Note: Numbers reflect the impacts of extending the number of products on the ITA list ( million). 17

18 Table A3.8 Optical, photographic and other instruments (Product 90) Country producer surplus consumer surplus tariff revenue Net change in welfare Country group Australia ITA Austria EU Belgium EU Brazil Non-ITA WTO Canada ITA Switzerland ITA Chile Non-ITA WTO China ITA Germany EU Spain EU Finland EU France EU Hong Kong ITA Hungary EU Indonesia ITA India ITA Israel ITA Italy EU Japan ITA Korea ITA Mexico Non-ITA WTO Malaysia ITA Netherlands EU Norway ITA Other EU EU OTHER NON-EU ITA ITA OTHER WTO, NON-ITA Non-ITA WTO Philippines ITA Rest of the world ROW Russia ROW Singapore ITA Thailand ITA Turkey ITA United Kingdom EU USA ITA Note: Numbers reflect the impacts of extending the number of products on the ITA list ( million). 18

19 APPENDIX 4 GRAVITY MODEL RESULTS We run gravity model estimations for the time period 1996 to 2005 on trade in ITA products under the two-digit HS85 and two-digit HS90 as well as for all ITA products as a group. Since the focus is on ITA countries barriers to trade, we exclude intra-eu trade flows from the study. Here, bilateral import to country i from country j in sector k at time t will depend on time-varying sector-specific bilateral tariffs (lnt), a set of time-invariant bilateral gravity factors including border, language and distance (gravity), time-varying importer GDP and GDP per capita (GDP and GDP/pop), time-invariant importer dummies (D i) and, finally, time-varying exporter dummies (D jt) and a time trend (t). The traditional gravity specification is: ln M ijkt = f (lnt ijkt, gravity ij, GDP it, GDP it / pop it, D, D i jt, t) The results from the gravity specification can be seen in the first column in the tables underneath. Table A4.1 shows the results for all the ITA products as a group, Table A4.2 shows the results for the ITA products under HS85 and Table A4.3 shows the results for the ITA products under HS90. We find that all the gravity variables generally enter with the expected sign: low tariffs, common language, common border and a short distance have a positive impact on trade. As expected, the elasticity of import with respect to importer s GDP is around one, whereas the elasticity of import with respect to GDP per capita unexpectedly turns out to be negative. The gravity model has typically formed the basis for a quantification of NTBs by interpreting the residual (the difference between actual and predicted trade) as missing trade between two trading partners. 5 This is the residual approach. One important criticism of the residual approach is that the residual captures a broad range of factors where some of these factors are trade-related but where others are rooted in structural, institutional and regulatory factors in the domestic economy. To get a more precise estimate of how open an importing country is, we use the importer dummy variables approach as was developed in European Commission (2010). 6 Since the gravity model controls for factors that affect trade and that vary over time by including the importer country s GDP and GDP/capita in the gravity model, the importer dummies will capture importer-specific non-tariff factors that have an impact on trade and that do not vary over time. If the importer dummy is negative, it means that there are barriers to ITA import that hinder foreign exporters from penetrating the market. The size of the dummy variable will thus be more directly related to trade barriers than the residual in the traditional gravity model. The dummy variables are reported in the main report and are therefore not shown here. 5 The gravity model approach was originally developed by Leamer (1998). See also Beghin and Bureau (2001) for a discussion of the various methods to quantify NTMs and Harrigan et al. (2003) for an application of the gravity model for Japan. 6 See European Commission (2010), Barriers to Trade between the EU and Japan, a study by Copenhagen Economics on behalf of DG Trade. 19

20 Table A4.1 Gravity regression results for all ITA products VARIABLES Gravity Gravity and importer RCA Gravity and importer and exporter RCA Gravity, RCAs and SDOCs Gravity, RCAs and ITA SDOCs Tariffs *** *** *** *** *** [-60.88] [-43.64] [-33.83] [-33.70] [-33.90] Distance *** *** *** *** *** [-67.67] [-39.63] [-30.52] [-30.50] [-30.59] Language 0.808*** 0.765*** 0.894*** 0.900*** 0.897*** [29.72] [17.81] [15.16] [15.24] [15.20] Border 0.430*** 0.571*** 0.890*** 0.878*** 0.864*** [7.517] [7.305] [9.189] [9.045] [8.910] Importer GDP 0.724*** 0.737*** 0.758*** 0.761*** 0.763*** [93.51] [37.26] [32.11] [32.14] [32.25] Importer GDP per capita *** *** *** *** *** [-3.867] [-8.812] [-4.793] [-5.282] [-5.375] Exporter GDP *** *** *** *** *** [-11.80] [-3.045] [-8.324] [-8.344] [-8.243] Exporter GDP per capita 0.401*** 0.385*** 0.399*** 0.396*** 0.390*** [32.50] [23.48] [8.046] [8.001] [7.854] Importer R&D expenditure 0.092*** 0.088*** 0.092*** 0.093*** [3.556] [2.851] [2.971] [3.022] Importer internet access 0.101*** 0.083*** 0.112*** 0.113*** [4.314] [2.877] [3.663] [3.700] Importer resident patent applications 0.090*** 0.099*** 0.096*** 0.096*** [5.935] [5.430] [5.278] [5.282] Exporter R&D expenditure 0.449*** 0.451*** 0.446*** [9.513] [9.560] [9.435] Exporter internet access [0.496] [0.487] [0.596] Exporter resident patent applications 0.240*** 0.241*** 0.239*** [10.49] [10.50] [10.42] Importer CA Type CD [1.302] [1.352] Importer CA Type G 1.722*** 1.714*** [5.444] [5.422] Importer CA Type Å 0.314** 0.315** Importer CA Type F 0.170** [2.003] [2.007] [2.427] Importer CA Type F in ITA countries 0.228*** [3.148] Importer CA Type F in non-ita countries [-1.376] Constant * ** ** [1.247] [0.956] [1.851] [2.113] [2.164] Observations R-squared

21 Table A4.2 Gravity regression results for ITA products under HS85 VARIABLES Gravity Gravity and importer RCA Gravity and importer and exporter RCA Gravity, RCAs and SDOCs Gravity, RCAs and ITA SDOCs Tariffs *** *** *** *** *** [-29.58] [-22.17] [-12.99] [-12.71] [-12.92] Distance *** *** *** *** *** [-69.50] [-43.68] [-41.39] [-41.58] [-41.67] Language 1.127*** 1.084*** 1.158*** 1.169*** 1.169*** [32.12] [21.73] [19.28] [19.38] [19.41] Border 0.475*** 0.670*** 1.020*** 0.983*** 0.974*** [6.003] [6.670] [8.823] [8.524] [8.470] Importer GDP 0.879*** 0.889*** 0.930*** 0.935*** 0.936*** [81.13] [33.41] [31.59] [31.75] [31.83] Importer GDP per capita *** *** *** *** *** [-6.118] [-10.33] [-5.628] [-6.888] [-6.950] Exporter GDP *** *** *** *** *** [-15.53] [-5.678] [-8.553] [-8.664] [-8.592] Exporter GDP per capita 0.413*** 0.378*** 0.165** 0.160** 0.156** [22.12] [15.77] [2.372] [2.305] [2.240] Importer R&D expenditure 0.134*** 0.130*** 0.141*** 0.143*** [3.903] [3.488] [3.810] [3.857] Importer internet access 0.179*** 0.138*** 0.204*** 0.205*** [6.177] [4.290] [5.981] [6.010] Importer resident parent applications 0.077*** 0.082*** 0.079*** 0.079*** [3.761] [3.602] [3.468] [3.471] Exporter R&D expenditure 0.726*** 0.730*** 0.727*** [10.34] [10.40] [10.38] Exporter internet access 0.151*** 0.150*** 0.151*** [2.915] [2.895] [2.936] Exporter resident patent applications 0.295*** 0.296*** 0.294*** [9.259] [9.276] [9.240] Importer CA Type CD 0.254** 0.258** Importer CA Type F 0.479*** [2.316] [2.351] [6.393] Importer CA Type G 2.623*** 2.634*** [7.600] [7.643] Importer CA Type Å 0.748*** 0.747*** [3.172] [3.169] Importer CA Type F in ITA countries 0.529*** [6.973] Importer CA Type F in non-ita countries 0.280** [2.205] Constant *** *** *** [-0.461] [-0.155] [5.493] [6.327] [6.378] Observations R-squared

22 Table A4.3 Gravity regression results for ITA products under HS90 VARIABLES Gravity Gravity and importer RCA Gravity and importer and exporter RCA Gravity, RCAs and SDOCs Gravity, RCAs and ITA SDOCs Tariffs *** *** *** *** *** [-16.62] [-11.36] [-8.593] [-8.172] [-8.341] Distance *** *** *** *** *** [-42.10] [-28.80] [-27.84] [-27.79] [-27.94] Language 0.599*** 0.570*** 0.878*** 0.882*** 0.881*** [14.04] [10.20] [14.74] [14.75] [14.74] Border 0.512*** 0.663*** 0.952*** 0.936*** 0.921*** [5.622] [5.742] [7.587] [7.460] [7.366] Importer GDP 0.907*** 0.947*** 0.968*** 0.977*** 0.980*** [72.45] [29.40] [27.51] [27.70] [27.84] Importer GDP per capita *** *** *** *** *** [-4.196] [-7.297] [-5.907] [-6.678] [-6.756] Exporter GDP 0.072*** 0.148*** * * * [2.917] [4.470] [-1.765] [-1.806] [-1.827] Exporter GDP per capita 0.501*** 0.546*** [16.89] [14.74] [1.027] [0.980] [0.897] Importer R&D expenditure 0.161*** 0.112*** 0.120*** 0.123*** [4.174] [2.771] [2.946] [3.039] Importer internet access 0.150*** 0.153*** 0.199*** 0.200*** [4.330] [4.109] [5.057] [5.101] Importer resident parent applications 0.128*** 0.140*** 0.142*** 0.141*** [5.475] [5.454] [5.516] [5.492] Exporter R&D expenditure 0.828*** 0.832*** 0.830*** [7.685] [7.737] [7.720] Exporter internet access 0.254*** 0.253*** 0.259*** [3.310] [3.300] [3.388] Exporter resident patent applications 0.353*** 0.354*** 0.354*** [7.089] [7.117] [7.119] Importer CA Type CD [1.508] [1.571] Importer CA Type G 1.174** 1.165** [2.504] [2.488] Importer CA Type Å Importer CA Type F 0.292*** [ ] [ ] [3.679] Importer CA Type F in ITA countries 0.345*** [4.308] Importer CA Type F in non-ita countries [-0.257] Constant ** *** *** *** [1.531] [1.986] [2.732] [2.966] [3.008] Observations R-squared

23 Extending the gravity model framework: comparative advantages In the second and third column we extend the traditional gravity model by including a set of ITA specific variables that make the gravity model more applicable to this set of products: R&D expenditures as a measure of technological strength If the importing country has a high level of R&D expenditures we expect the country to demand less ITA products since its degree of self-sufficiency is higher. If the exporting country has a high level of R&D expenditure we expect the country to have a comparative advantage in highly sophisticated products, and we therefore expect a positive impact on trade. Internet access as a measure of technological advancement If the importing country has good internet access we expect the country to have a high level of technological advancement and the country will therefore have a higher demand for ITA products. If there is poor internet coverage, however, demand for ITA products will be lower. Resident patent applications as a measure of competitiveness in ITA products If the importing has a high number of patent applications by its resident companies there will be less of a need to import ITA products. By symmetry, countries with a large number of patent applications will be more competitive and will be able to export more We find that all of the proxies for importer and exporter comparative advantage in ITA products appear to be positive and enter with the expected sign. Extending the gravity model framework: conformity assessment Conformity assessment has been identified as one of the main NTBs to expanded trade in ITA products. We use information about the types of conformity assessment systems in place in ITA countries (see Table A4.4 and Table A4.5) to construct conformity assessment dummies that will allow us to identify trade impacts of introducing a particular system: Importer CA Type T AB: Dummy variable that takes the value 1 from the time when the importing countries implemented a system of certification by a regulator or by a third party and zero otherwise (type A or type B). Importer CA Type T CD: Dummy variable that takes the value 1 from the time when the importing countries implemented a system with an SDoC type C and type D and zero otherwise. Importer CA Type T E: E Dummy variable that takes the value 1 from the time when the importing countries implemented a system with an SDoC type E and zero otherwise. 23

24 Importer CA Type T F: F Dummy variable that takes the value 1 from the time when the importing countries implemented a system with an SDoC type F and zero otherwise. Importer CA Type T G: Dummy variable that takes the value 1 from the time when the importing countries implemented an SDoC system with no mandatory certification and zero otherwise (type G). Since none of the countries for which we have information have an SDoC system characterised as type E this dummy is excluded from the analysis. When we do not have update information about when the exact system was implemented, we take the conservative stance and assume that the system has been in place throughout the period. Table A4.4 Description of conformity assessment systems in place in ITA countries Type Short name Description CA Type A Regulator Certification by a regulator or delegated entity the equipment has to be submitted to the regulator or its delegated entity for certification. CA Type B 3rd party CAB Certification by 3 rd party the equipment has to be submitted to certification bodies recognised (or approved) by the regulator for certification. CA Type C SDoC Lab Test Reg Supplier's Declaration of Conformity (SDoC) type 1 the supplier or manufacturer of the equipment declares the equipment meets the technical and administrative requirement. A testing laboratory recognized by the regulator tests the equipment and the supplier registers this equipment with the regulator. CA Type D SDoC Lab Test No Reg Supplier's Declaration of Conformity (SDoC) type 2 - the supplier or manufacturer of the equipment declares the equipment meets the technical and administrative requirements on the basis of test reports by a testing laboratory recognized by the regulator. No registration of the equipment with the regulator is required. CA Type E SDoC Lab Reg No Test Supplier's Declaration of Conformity (SDoC) type 3 the supplier or manufacturer of the equipment declares the equipment meets the technical and administrative requirement. The supplier registers the equipment with the regulator. Testing of the equipment by recognized testing laboratory is not mandatory. If testing is undertaken, the choice of the testing laboratory rests with supplier or manufacturer. CA Type F SDoC No Test Supplier's Declaration of Conformity (SDoC) type 4 the supplier or manufacturer of the equipment declares the equipment meets the technical and administrative requirement. Registration of the equipment with the regulator is not required and testing of the equipment by recognized testing laboratory is not mandatory. If testing is undertaken, the choice of the testing laboratory rests with supplier or manufacturer. CA Type G No mandatory No mandatory assessment procedure In order to avoid multicollinearity with the constant term we exclude the Importer CA Type AB dummy. This means that the other CA dummies should be interpreted relative to the Importer CA Type AB dummy. We expect that the restrictiveness of the CA system is decreasing so that a CA Type CD is less restrictive than the CA Type AB. This means that the signs of the CA dummies are expected to be positive, which is also found to be the case. We are particularly interested in the CA Type F since this is the conformity assessment system adopted by the EU countries. We test if there is a difference in the impact on ITA and non- ITA countries. We do so by interacting the Importer CA Type F dummy with an ITA dummy variable, which takes the value 1 when the exporter is an ITA country and zero otherwise. We find that the impact on non-ita countries is significantly smaller than the impact on ITA countries. 24

25 Table A4.5 Conformity assessment systems used by ITA participants Type of conformity assessment procedure Country Type name adopted Australia F SDoC No Test Canada F SDoC No Test China B 3rd party CAB Dominican Republic F SDoC No Test El Salvador F SDoC No Test European Communities F SDoC No Test Hong Kong, China G No mandatory Honduras B 3rd party CAB India B 3rd party CAB Japan D SDoC Lab Test No Reg Jordan C SDoC Lab Test Reg Korea A Regulator Macao, China A Regulator Malaysia G No mandatory Mauritius B 3rd party CAB New Zealand F SDoC No Test Norway F SDoC No Test Philippines G No mandatory Taiwan A Regulator Singapore B 3rd party CAB Switzerland D SDoC Lab Test No Reg Thailand G No mandatory Turkey F SDoC No Test United States D SDoC Lab Test No Reg Source: WTO (2007) Draft List of the Types of Conformity Assessment Procedures for EMC/EMI Used by ITA participants. Barriers to trade in ITA products After accounting for tariffs, distance, language and macroeconomic factors we find that there are significant barriers to ITA trade in most import countries, cf. Figure A4.1. Only Malaysia turns out to have no or very few restrictions on ITA imports. Hong Kong, Singapore, Taiwan and Malaysia do not seem to impose significant barriers on ITA imports. 25

26 Figure A4.1 General restrictiveness of ITA import Size of NTBs on all ITA products Average restrictiveness is Note: Restrictiveness is calculated after accounting for other variables (language, distance, tariffs and macroeconomic factors). It is measured by the importer dummy from the gravity model for all ITA goods. The importer dummies take the value 1 for a particular importer country and 0 otherwise. See Appendix 2 for further details on the gravity model. The importer dummy in Hong Kong, Singapore, Taiwan and Malaysia do not turn out significant in the gravity model. The Malaysian importer dummy is positive which means that Malaysia imports more than predicted by the gravity model. Source: Copenhagen Economics gravity model for ITA goods. We find that the average restrictiveness of imports in ITA products under HS85 (equal to 1.39) is larger than the restrictiveness on ITA products as a whole (equal to 1.11). Of the larger countries, Russia, Japan, Turkey and Indonesia appear to be very restrictive, cf. Figure A4.2. The EU countries only come in 15 th. 26

27 Figure A4.2 Restrictiveness of import of ITA products under HS85 Size of NTMs on ITA products under HS85 Average restrictiveness is Note: Restrictiveness is calculated after accounting for other variables (language, distance, tariffs and macroeconomic factors). It is measured by the importer dummy from the gravity model for ITA goods under HS85. The importer dummies take the value 1 for a particular importer country and 0 otherwise. The sign of the importer dummies has been reversed for expositional purposes. See Appendix 4 for further details on the gravity model. Source: Copenhagen Economics gravity model for ITA goods. Barriers to trade in ITA products under HS90 (average importer dummy is 0.61) are on average smaller than barriers to trade in HS85 products. Again, Russia, Turkey and Indonesia turn of to have large barriers to import of ITA products, cf. A4.3. A large number of countries do not seem to impose significant barriers to import of ITA products under HS90 (importer dummies turn out insignificant). 27

28 Figure A4.3 Restrictiveness of import of ITA products under HS90 Size of NTBs on ITA products under HS90 Average restrictiveness is Note: Restrictiveness is calculated after accounting for other variables (language, distance, tariffs and macroeconomic factors). It is measured by the importer dummy from the gravity model for ITA goods under HS90. The importer dummies take the value 1 for a particular importer country and 0 otherwise. The sign of the importer dummies has been reversed for expositional purposes. See Appendix 4 for further details on the gravity model. Source: Copenhagen Economics gravity model for ITA goods. 28

29 APPENDIX 5 NTBS IN THE EC MARKET ACCESS DATA BASE Table A5.1 NTBs on electronics Type of barrier Export country Description of problem Standards and Other Technical Requirements US Most electrical products in the EU go through a product approval process called "internal production control", which in international discussions often is referred to as Suppliers Declaration of Conformity (SDoC). The European Commission requested that the US Occupational Safety & Health Administration (OSHA) deregulates its current procedures that require products to go through nationally recognised testing laboratories, ideally by a move towards SDoC. Nationally Recognised Testing laboratories (NRTLs) are third-party laboratories that have met OSHA requirements for performing safety testing and certification of electrical and other products used in the workplace. NRTLs test and certify these products to determine whether they conform to appropriate U.S. product-safety testing standards. SDoC, applicable for most electrical products placed on the European Union market, obliges manufacturers to adhere to strict safety requirements and obliges them to be able to document compliance at all times. It leaves however the detailed modalities for the proof of compliance to the manufacturer and does not require him to use a locally recognised test laboratory. They therefore are free to use the services of any competent (e.g. accredited) test laboratory or use in-house competence. Other Non-Tariff Measures Legislation on Patents (Including Plant Varieties) Korea Russia Taiwan On 20 October 2008 OSHA published a request for information and comments on a proposal to permit the use of Suppliers Declaration of Conformity (SDoC) as an alternative to the Nationally Recognised Testing laboratories (NRTLs) product-approval process. The public has had 90 days to respond (i.e. until 20 January 2009). There is a general non-recognition of international standards and compliance certification. Local certification process takes about 1 to 2 months which seems to be clearly too long e.g. in comparison to the selfdeclaration system adopted in the EU. Government procurement seems open to foreign companies but in practice, industry is required to produce local standards certification, making it harder for foreign brands to qualify. Russia classifies all encryption products (including mobile phones) as sensitive risk products for which an expensive and very burdensome licensing procedure needs to be followed. The issue is under discussion in the framework of Russias WTO accession negotiations. The complaint concerns alleged trade barriers maintained by Chinese Taipei by which it granted compulsory licences in respect of certain patents held by Philips for CD-R technology. It is alleged that these measures are inconsistent with Articles 28 and 31 of the TRIPs Agreement. The complaint was lodged on 15 January 2007, and the Notice of Initiation of an examination procedure was published on 1/3/2007 (O.J. C 47/ ). The Commission reported to the TBR Committee on 11 January 2008 and communicated the findings to Chinese Taipei and the Complainant on 29 January 2008 in the form of an executive summary and the full report. The investigation focussed on the compulsory licence provisions of the Patent Act of Chinese Taipei and specific decisions taken under that Act to grant compulsory licences over patents owned by Philips. The Commission's TBR report found violations of the TRIPs Agreement stemming from first, the Patent Act of Taiwan and second, from the manner in which the Act had been applied in respect of patents for CD-Rs held by Philips. The Commission services concluded that the provision of the Patent Act dealing with compulsory licences was inconsistent with Article 28 of the TRIPs Agreement, because it allows the grant of the such licences where there is no more than a refusal to deal on the part of the patent owner. Further, the Commission services concluded that the interpretation of various procedural requirements relating to the grant of compulsory li- 29