Back Matter
  • 1, International Monetary Fund

Appendix I

The GTAP Model

The Global Trade Analysis Project (GTAP) model used in this paper is a comparative static, global general equilibrium model based on neoclassical theory.36 Firms maximize their profits while consumers maximize their utility. All markets are assumed to be perfectly competitive, and constant returns to scale prevail in all production and trading activities.

Firms use both a composite of primary factors and a composite of intermediates to produce their output according to Leontief production technology. The primary factor composite is a constant elasticity of substitution (CES) function of labor, capital, land and natural resources, while the intermediate composite is a Leontief function of material inputs, which are in turn CES blends of domestically produced goods and imports. Imports are sourced from all regions, with their share depending on trading prices (the Armington approach).

On the demand side, each country or region is assumed to have a “super” household disposing of regional income in fixed proportions in the form of private consumption, government expenditure and savings. Household consumption is assumed to be a constant difference in elasticities (CDE) function of various consumer goods while government expenditure is based on a CES function of various commodities. Both household and government consumption are CES blends of domestically produced goods and imports, which are in turn sourced from all trading regions based on the Armington approach.

In closing the model, regional savings are assumed to be homogenous and contribute to a global pool of savings, which is then allocated among regions for investment in response to changes in regional expected rates of return. These changes are assumed to be equalized across regions, thus giving rise to capital (i.e., savings) mobility across regions. This allows for greater changes in the trade balance as a result of trade liberalization and tends to dampen the terms of trade effects. In contrast to savings, capital stocks are assumed to be immobile across regions, although they are perfectly mobile within a region, as is labor. Land and natural resources are industry-specific, and only limited transformation of their uses among industries is possible.

The simplicity of the GTAP model makes its simulation results relatively easy to interpret, but limits its capacity to deal with more complex economic issues, such as the adjustment path over time and long-term effects of trade policies associated with investment accumulation, technology and productivity change. Also absent in the model are adjustment costs associated with trade liberalization. These limitations must be kept in mind when interpreting the results presented in this paper.

The GTAP database provides data on key trade policies, as well as on other essential data for a large number of countries and commodities. The base year for the data is 1997. The region/country and commodity aggregations are shown in Table A1 and the elasticities used in the simulations are presented in Table A2.

Data Aggregation and Armington Elasticities

Table A1.

Country/Region and Industry Aggregations Used in the Model

article image
Table A2.

Central Scenario Elasticities of Substitution in Demand for Goods

article image
Source: Based GTAP database version 5.

Table B1.

Effects of MFA Quota Removal on T&C Trade, 2007 1/

(values in 1997 prices)

article image
Source: Simulations with the GTAP model as described in the text.

Results corresponding to Scenario 1 in Table 12.


  • Ahmed, Sadiq, and Zaidi Sattar, 2003, “Trade Liberalization, Growth, and Poverty Reduction: The Case of Bangladesh,” paper presented at the World Bank ABCDE Conference, Bangalore, July.

    • Search Google Scholar
    • Export Citation
  • Anderson, K., B. Dimaranan, T. Hertel, and W. Martin, 1996, “Asia-Pacific Food Markets and Trade in 2005: A Global, Economy-wide Perspective,” Seminar Paper 96-05, Center for International Economic Studies, University of Adelaide, South Australia.

    • Search Google Scholar
    • Export Citation
  • Bangladesh Enterprise Institute and World Bank, 2003, Improving the Investment Climate in Bangladesh (Washington: World Bank).

  • Bhattacharya, Debapriya, 2003, “Final Countdown of the MFA: Fallout for the LDCs,” (Dhaka: Center for Policy Dialogue).

  • Bhattacharya, Debapriya, and Mustafizur Rahman, 2000, “Experience with Implementation of WTO-ATC and Implications for Bangladesh,” (Dhaka: Center for Policy Dialogue).

    • Search Google Scholar
    • Export Citation
  • Center for Policy Dialogue, 2003, Phasing Out of the Apparel Quota: Addressing Livelihood Concerns (Dhaka: Center for Policy Dialogue and the University Press Ltd.).

    • Search Google Scholar
    • Export Citation
  • Cookson, Forest, 2003a, “A Short Note on Export Development,” Dhaka, March.

  • Cookson, Forest, 2003b, “Facing 2005—Outlook for the Garment Sector,” Dhaka, March.

  • Dean, Judith, 2002, “Removing Textile and Apparel Trade Barriers: The Impact on Developing Country Exporters” (unpublished; Washington: U.S. International Trade Commission).

    • Search Google Scholar
    • Export Citation
  • Dimaranan, B.V., and R.A. McDougall, 2002, Global Trade, Assistance, and Production: The GTAP 5 Data Base, Center for Global Trade Analysis, (West Lafayette, Indiana: Purdue University).

    • Search Google Scholar
    • Export Citation
  • Gereffi, G., 1999, “International Trade and Industrial Upgrading in the Apparel Commodity Chain,” Journal of International Economics, Vol. 48, No. 1, pp. 3770.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gherzi Textil and Project Promotion and Management Associates, 2002, “Strategic Development and Marketing Plans for RMG and Related Industries,” report submitted to the Ministry of Commerce, Government of Bangladesh, December.

    • Search Google Scholar
    • Export Citation
  • Harrison, W.J., and Pearson, K.R., 1996, “Computing Solutions for Large General Equilibrium Models Using GEMPACK,” Computational Economics, Vol. 9, pp. 83127.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hertel, T.W., ed., 1997, Global Trade Analysis Using the GTAP Model, (New York: Cambridge University Press).

  • Hummel, David, 2001, “Time as a Trade Barrier”, (unpublished; West Lafayette, Indiana: Purdue University) July.

  • Islam, Sadequl, 2001, The Textile and Clothing Industry in Bangladesh in a Changing World (Dhaka: Center for Policy Dialogue and the University Press, Ltd.).

    • Search Google Scholar
    • Export Citation
  • Krishna, K., R. Erzan, and L. Tan, 1994, “Rent Sharing in the Multi-Fibre Arrangement: Theory and Evidence from U.S. Apparel Imports from Hong Kong,” Review of International Economics, Vol. 2, No. 1, pp. 6273.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sattar, Zaidi, 2000, “Foreign Direct Investment in Bangladesh: Issues of Long-term Sustainability,” Policy Note, World Bank Resident Mission, Dhaka.

    • Search Google Scholar
    • Export Citation
  • Spinanger, Dean, and Samar Verma, 2003, “The Coming Death of the ATC and China’s WTO Accession: Will Push Come to Shove for Indian T&C Exports?” (unpublished, Kiel and New Delhi: Institute for World Economics and OXFAM), June.

    • Search Google Scholar
    • Export Citation
  • Spinanger, Dean, and Jan Peter Wogart, 2001, “Will the Emperor Wear Clothes from Bangladesh in 2005?paper presented at the IEA Meeting in New Orleans, January.

    • Search Google Scholar
    • Export Citation
  • United States International Trade Commission (USITC), 2003, U.S. Trade and Investment with Sub-Saharan Africa, Third Annual Report, Investigation No. 332-415, Publication 3650 (December).

    • Search Google Scholar
    • Export Citation
  • United States International Trade Commission (USITC), 2004, Textiles and Apparel: Assessment of the Competitiveness of Certain Foreign Suppliers to the U.S. Market, USITC Publication No. 3671 (Washington: USITC), January.

    • Search Google Scholar
    • Export Citation
  • World Bank, 2004, “Implications for Pakistan of Abolishing Textile and Clothing Export Quotas,” mimeo, April 17.

  • World Trade Organization, 2000, Trade Policy Review: Bangladesh, Report by the Secretariat (Geneva: World Trade Organization).

  • Yang, Y., W. Martin, and K. Yanagishima, 1997, “Evaluating the Benefits of Abolishing the MFA in the Uruguay Round Package,” in Global Trade Analysis: Using the GTAP Model, ed. by Thomas Hertel, (New York: Cambridge University Press), pp. 25379.

    • Search Google Scholar
    • Export Citation

The authors wish to thank, without implication, Peter Berezin, Nur Calika, Hans Peter Lankes, Olin Liu, Jeremy Mark, Will Martin, Caroyl Miller, Lynge Nielsen, Zaidi Sattar, Dean Spinanger, and Marijn Verhoeven for their helpful comments, and Ross Arnold, Vlad Manole, Abul Quasem, and Dustin Smith for their valuable assistance with data collection. Mary Jo Marquez provided excellent editorial support.


See Dean (2002) for a concise summary of how MFA quotas have affected trade patterns across exporting countries.


See the IMF Board paper “Market Access for Developing Country Exports—Selected Issues,” SM/02/280,


In Bangladeshi usage, the term “RMG” covers items that are not necessarily garments (clothes), such as towels, napkins, etc. In this paper, the terms “T&C” and “RMG” are used interchangeably, unless the context demands otherwise.


For a fuller description of the Bangladeshi garment industry, see Bhattacharya (2003), Bhattacharya and Rahman (2000), and Islam (2001).


Despite large imports, the country ran a trade surplus of US$2.8 billion in the combined textile and clothing sector.


AGOA provides exclusive quota- and duty-free market access for exports from eligible countries in Sub-Saharan Africa. According to USITC (2003), U.S. imports of textile and garments from AGOA-eligible countries have indeed increased rapidly since 2001. It is not clear, however, whether this has been at the expense of Bangladeshi exports.


Norway’s quotas have been removed for all exporting countries, while tariffs on exports from LDCs were eliminated in 2002.


If both a quota and tariff are imposed on a product, only one of them can be binding. If the quota is binding, the tariff simply allows the importing country to recoup some of the quota rent that would otherwise accrue to the exporting country and the entire preference margin for Bangladesh is the tariff equivalent of the quota. If the quota is not binding, the entire preference margin for Bangladesh is the level of the tariff.


Under the EU’s so-called “SAARC cumulation” rules, Bangladeshi products made with inputs from the South Asian Association for Regional Cooperation (SAARC) region (including Bangladesh, Bhutan, India, the Maldives, Nepal, Pakistan, and Sri Lanka) are eligible for preferential treatment under the EBA, if they meet the minimum value-added requirement.


The closure of land ports for textile imports from India in March 2002 has increased difficulties for the knitwear industry. This policy is supposed to limit smuggling, which the local textile industry claims to be a major impediment to its efficiency. The motivation for the policy is unclear.


See BNA, Inc., International Trade, Vol. 21, No. 12, March 18, 2004, p. 496.


For an explanation of how quota removal would affect unrestricted markets, see Yang and others (1997).


The Finger-Kreinin similarity index is defined as IAB=i=1nMin(SiA,SiB)100, where SiA is the share of product i in country A’s exports to an export market, SiB is the share of product i in country B’s exports to the same market, and n is the number of products. It should be noted that the index is sensitive to the level of product disaggregation. The more disaggregated the products are for the same level of aggregate exports, the lower the index is. In this study, the index is computed at the quota level.


Under the MFA and ATC, exporting countries are allowed some flexibility in using some of their base quotas in terms of the so-called carry-forward, carry-over, and swings (shifting a specified proportion of a quota category to another within the same year). Actual annual growth rates of quotas therefore may differ from base rates.


For the ease of exposition, it is assumed here that all exporting countries produce a homogeneous product. However, the story would be similar even if products are assumed to be differentiated by country of origin.


This assumes that the effects of the EBA have fully worked through before the quota removal in 2005. This conclusion also depends on the existence of barriers (e.g., rules of origin and infrastructure bottlenecks) that prevent costless re-direction of Bangladeshi exports from the U.S. to the EU market—otherwise, for a small country like Bangladesh, U.S. quotas would become redundant immediately after the introduction of the EBA. However, the prices of U.S. quotas on Bangladeshi exports should have been subject to downward pressure as long as the EBA has had the effect of easing barriers against Bangladeshi exports.


The demand elasticity depends importantly on the elasticity of substitution between exports from Bangladesh and its competitors.


Note that although these numbers are also computed at the quota level they are not comparable with those for the U.S. markets as product classifications differ. All results are based on data provided by the European Commission.


See Appendix Table A. 8 in World Bank (2004). These ETEs seem to be extraordinarily high, ranging from 82 percent to several hundred percent.


Data on Indian ETEs at the product level in the EU market are not available.


The similarity index failed to capture the potential export competition because binding quotas on one country raise exports from its competitors while reducing that country’s exports. The resulting low export similarity resembles the endogeneity problem of calculating trade-weighted average tariff.


Permission is usually given only for manufacturers that have at least one backward linkage. For a more detailed discussion of the FDI regime in Bangladesh, see Sattar (2000).


A buying house is an intermediary without its own retail operations.


Firms may be allowed to import up to four months’ needs if the proposed export complies with norms under the Import Policy Order, notably minimum value addition of 25 percent. This generally favors only large firms.


The rate has been progressively reduced from 25 percent in FY01, and is expected to be completely phased out by June 2006.


Estimates based on the impact of quota removal to date, performance in 10 key garment categories and discussions with industry leaders.


It is assumed in the model that quota rents accrue to the exporting country. There is some evidence, however, that quota rents are shared between the exporting and importing countries (Krishna and others, 1994).


At the margin, this is the unlimited labor supply model. Given the high unemployment rate and the vast pool of the underemployed in rural Bangladesh, this assumption is not unreasonable. Most existing studies tend to assume fixed labor supply, which may not accurately reflect the fact that layoffs are common, except in the public service sector. The numeraire of the model is the average global factor price.


The results reported are based on the average (8 percent) of the estimates of quota premiums in Bangladesh. See Appendix II for results on T&C exports for other countries/regions included in the model.


In the model products are differentiated by country of origin (the Armington assumption). There are two sets of elasticities of substitution: those between domestic products and imports and those between imports by country of origin (see Appendix Table A2 for the elasticity values).


Even within textile and clothing exports, the product range is limited: eighty percent of exports of knitwear products are men’s and boys’ shirts, and T-shirts, while 75 percent of woven products are men’s and boys’ shirts and trousers and women’s and girl’s trousers. While these products are not separately modeled, this concentration could bring additional vulnerability to external competition.


The long-term adverse effect of reduced investment on the balance of payments and growth is not captured here as investment does not affect the capital stock given the static nature of the model.


The United States has already reimposed quotas on three categories of textile imports from China in December 2003 and is reportedly trying to negotiate a more broadly based new quota system to forestall the expected rapid expansion of Chinese exports after 2004.


See Centre for Policy Dialogue (2003), for a discussion of the potential social impact of the quota removal.


Full documentation of the GTAP model and its companying database can be found in Hertel (1997) and Dimaranan and McDougall (2002). The GTAP model is solved using the software GEMPACK (Harrison and Pearson 1996).

The End of Textiles Quotas: A Case Study of the Impacton Bangladesh
Author: Mr. Montfort Mlachila and Yongzheng Yang