Journal Issue

Pakistan: Selected Issues and Statistical Appendix

International Monetary Fund
Published Date:
January 2001
  • ShareShare
Show Summary Details

III. Perspectives on Pakistan’s Export Performance67

A. Introduction

136. Following rapid export growth from the mid-1980s to the early 1990s, Pakistan’s merchandise export performance has been disappointing in recent years. With stagnating exports, import financing became more difficult, and the continued accumulation of external debt and pressures on foreign exchange reserves became prominent features of balance of payments developments. The sluggish export performance is often explained with the congruous weakening in the performance of Pakistan’s cotton and textile sectors. The latter increasingly dominated export developments during the last two decades, as growth in the tradables sector appears to have been based largely on the expanding downstream processing of the domestic cotton output.

137. This Section attempts to ascertain the factors explaining the recent stagnation in merchandise exports. It builds on the analysis of Pakistan’s export performance during the 1980s and 1990s, especially with regard to the linkages between performance and the structure of exports (commodity composition and geographical destination of exports). It also attempts to put Pakistan’s trade performance into perspective by comparing it with the export performance and structure, and trade regime of important regional competitors. In terms of methodology, the analysis relies on descriptive statistics in the trade domain and on the method of constant market shares analysis. The latter allows one to study the role of three key forces in the export performance of a country: the overall demand for imports in trading partner countries; the commodity composition of trading partners’ import demand; and the responsiveness of the export supply to changes in the commodity composition in the import demand of trading partners.

138. The Section is organized as follows: subsection B documents trends in overall export performance, the commodity composition of exports, and the geographical direction of trade. Subsection C compares Pakistan’s export performance and trade regime with those of important competitors in the region, including Bangladesh, China, India, Indonesia, Sri Lanka, and Thailand. Subsection D describes the technique of constant market share analysis and the results obtained from applying it to four important destinations of Pakistan’s exports. Subsection E concludes and offers policy recommendations.

B. Pakistan’s Export Performance, 1980–200068

Overall export performance

139. Pakistan’s export earnings in value terms fluctuated dramatically during the last two decades (Chart III-1). Nevertheless, a decomposition of exports into a medium-term “trend” component and a cyclical-irregular component shows a clearly recognizable pattern of a sustained acceleration in export growth rates from the mid to late 1980s and a decrease in export earnings growth from 1990.69

Chart III-1.Pakistan: Export Performance, 1980 - 2000 1

Source: Staff calculations and data provided by Pakistani authorities.

1/ Fiscal year basis; Fiscal year runs from July 1 to June 30.

140. Chart III-1 indicates that in the mid-1980s, export volume growth was the dominant force behind the increase in the dollar value of exports. The improved export performance during this period reflected the (a) adoption of more flexible exchange rate management coupled with the launch of a trade liberalization program (Box III-1); (b) promotion of private sector investment; (c) expansion of output in cotton and textiles; and (d) the adoption of direct export subsidies. The decline in the growth of export values during the 1990s reflected both a decrease in volume and unit value growth rates, although the contribution of the two components varied considerably. During the early to mid-1990s, increases in unit values were the main factor behind export value growth. From the mid-1990s, a decline in export volumes dominated developments. As argued below, the deceleration in the growth rate of export volumes was largely related to negative productivity shocks in the agricultural sector, which in turn affected output in related downstream industries.

141. In 1999, exports were hit by the East Asian financial crisis, the recession in Japan, and the slowdown in both world trade and output in 1998 more generally. In 2000, export volumes rebounded with the bumper cotton crop, which provided a boost to the textile industry. Nominal export values rose somewhat less, as world prices for cotton had fallen significantly during the calendar year 1999.

Commodity composition of exports

142. The expanding downstream processing of the domestic cotton crop increasingly dominated Pakistan’s export developments during the last 20 years and was reflected in a dramatic change in the commodity composition of exports. Primary commodities—mostly cereal, cereal-based products, and raw cotton—as a share in total exports decreased from about 37 percent in the early 1980s to roughly 13 percent in 2000. At the same time, the share of textile manufactures—including yarn, cloth and other textile fabrics, apparel, and clothing accessories—more than doubled from about 35 percent in 1980 to about 75 percent in 1998. The shares of most other manufacturing categories except for miscellaneous manufacturing items were small and, in some cases, even decreased over the last 20 years, as shown in the first panel in Chart III-2 and Table III-1 (which are based on export data at the two-digit code level of the Standard International Trade Classification (SITC)).

Box III-1.Trade Liberalization in Pakistan Since the mid-1980s

Pakistan’s international trade regime has been liberalized over the past 15 years. Tariffs have been reduced, nontariff restrictions have been eased, and the role of special exemptions has declined. Nevertheless, import tariffs in Pakistan remain higher than those in most economies in the region, and restrictions are still in place on the import and export of a number of items.

Import tariffs have been reduced markedly since the late 1980s. The maximum tariff rate was cut in stages from 225 percent in 1988 to 35 percent in 1999. The simple average tariff rate declined from 123.5 percent in 1986 to 23.8 percent in 1999. The number of tariff slabs, or rates, has also been reduced, from 10 in 1986 to 5 at present. In addition, paratariffs, which include surcharges, fees, and regulatory duties have been reduced; paratariffs are now integrated into the statutory tariff schedule.1 These changes have had a major impact on customs duty receipts, which have declined from over 6 percent of GDP, on average, in the late 1980s to 2 percent of GDP in 1999/2000.

Progress has also been made in reducing nontariff barriers. Until the early 1980s, Pakistan relied on a trade system that specified permitted imports (positive list). This was replaced with a negative list of imports; import of all items not on the negative list was permitted from 1983/84. The negative list has been narrowed from over 500 items in the mid 1980s to 44 items at present. The items that remain on the negative import list include cotton fabrics, carpets, bed linen, apparel, and clothing accessories.2 In addition, imports of some 50 products, including vehicles, are subject to administrative procedures. On the export side, the number of items subject to quantitative restrictions has also been reduced, although 11 items remain banned and 13 products—including wheat and its milled products, cotton, rice, metals, and urea fertilizers—are subject to specific conditions and procedures.

The role of special import duty concessions and exemptions has declined in recent years. The system of statutory import tariffs in Pakistan is modified by a series of end-user-based concessions and exemptions, referred to as Statutory Regulatory Orders (SROs). The ratio of actual customs revenue from items covered by SROs to the revenue that would have been collected had the SROs not been in place has declined from close to 50 percent in 1996/97 to 36 percent in 1999/2000. In addition, the number of import-related SROs has been reduced.

Most imports continue to be subject to withholding taxes. The withholding tax rate was raised from 5 percent to 6 percent in July 2000.

In 1998, the Pakistan authorities agreed with the WTO to phase out quantitative restrictions on imports. The phasing-out schedule was subsequently suspended due to the weak balance of payments position.

Chart III-2.Pakistan: Structure of Exports by Commodity Groups

Source: Data from Trade Analysis and Reporting System, United Nations.

1/See text for details.

Table III-1.Pakistan: Shares in Total Exports by Commodity Groups 1/(In percent of total exports)
Textile Yarn and Fabrics37.842.650.151.852.049.5
Cereals and Cereal-Based Products14.
Miscellaneous Manufactured Goods2.
Leather and Leather Products, and Fur4.
Textile Fibers12.913.
Sugar and Sugar-based Products, and Honey1.
Fish and Fish-based Products2.
Instruments, Watches, and Clocks1.
Fruit and Vegetables1.
Crude Animal, Vegetable Material1.
Petroleum and Products4.
Medical Products0.
Metal Manufactures0.
Nonmetallic Mineral Manufactures0.
Oil Seeds, Nuts, and Kernels0.
Coffee, Tea, Cocoa, and Spices0.
Non-electrical Machinery0.
Metalliferous Ores, and Scrap0.
Fertilizerers and Minerals0.
Miscellaneous Manufactured Food Products0.
Transport Equipment0.
Electrical Machinery0.
Source: Data from the United Nations, Trade Analysis and Reporting System; and staff calculations.

Data shown for the top twenty five commodity shares ranked using average value during 1994—1998. Data points for 1980–85,1986-90, and 1991–1995 are averages for the five-year periods.

Source: Data from the United Nations, Trade Analysis and Reporting System; and staff calculations.

Data shown for the top twenty five commodity shares ranked using average value during 1994—1998. Data points for 1980–85,1986-90, and 1991–1995 are averages for the five-year periods.

143. The fact that one commodity category accounted, on average, for almost 75 percent of Pakistani exports during the last five years suggests a high degree of concentration. To measure the degree of concentration in the commodity composition of exports over time, Gini-Hirschman70 concentration indices were computed, using the shares of 59 different commodity groups (according to the disaggregation at the two digit SITC code level). The value of the Gini-Hirschman index is bound to be in the interval between 0 and 1. The closer the index value is to 1, the higher is the degree of concentration in the commodity composition. As shown in the lower panel of Chart III-2, the index increased from 0.43 in 1981 to 0.56 in 1998, confirming the hypothesis that the concentration in the commodity composition of exports was increasing during the last two decades.71 Data analysis also reveals that the rise in the index can be largely attributed to the increase in the share of textile-based exports.

144. The data illustrate how Pakistan’s transition from a primary commodity exporter to an exporter of manufactures was closely linked to the downstream processing of the domestic cotton crop. As noted in Box III-2, the local production of cotton provided for a natural competitive advantage to the development of the textile sector, although sector-specific incentives such as price distortions, subsidized loans, tax holidays, and protection also played a role.

145. The concentration on a single category of manufactures bears considerable risks, as it increases the vulnerability of exports developments to developments in a specific world market segment as well as to the vagaries of the domestic cotton production, given restrictions on the exports and imports of raw cotton that were in place until recently.72 As argued in Box III-2, a sequence of adverse productivity shocks to cotton production was an important if not the most important reason for the disappointing export performance in the area of textiles. In addition, world market prices for cotton and some cotton-based products stagnated during the second half of the 1990s and fell in 2000, which also limited the scope for export value growth.

Box III-2.Pakistan: Cotton and Textile Sector Developments, 1980–2000

The 1980s and early 1990s were the “golden” years of rapid growth in the cotton sector and the downstream processing of cotton in the textile industry, which in turn laid the basis for substantial increases in textile related exports. Cotton output tripled from 1980 to 1992, partly reflecting productivity increases due to better seeds and more use of pesticides (following the liberalization of the pesticide industry) and of fertilizer (following the partial liberalization of the industry).

This provided a natural supply base for the rapid development of the domestic textile industry from the early 1980s, which was reflected in a smaller gap between domestic cotton production and absorption by the industry. Growth in the industry was partly spurred by the increased demand for cotton garments in lieu of polyester-based textiles. Trade restrictions and price distortions resulting from government procurement and price setting kept domestic cotton prices below world market prices, which induced massive investment including from nonresidents, especially in yarn spinning and cloth weaving.1 Subsidized loans, tax holidays, and nontariff barriers including a ban on imports of industry-specific semi-manufactures further supported the development of the textile sector. With the expansion of the domestic industry, exports of raw cotton fell both in quantity and value.

From 1993, the cotton sector was hit by a series of negative productivity shocks, including leaf curl virus infestations during 1993–95 and the American bollworm infestation in 1996/97. This had detrimental effects on the textile industry, not only through the effects on the supply basis but also through the related effect on domestic producer prices. As trade restrictions limited imports and exports of cotton, and given the already small amounts of raw cotton exports, domestic wholesale prices for lint cotton began increasing at a faster pace than world market prices from 1992. As a result, the competitiveness of the textile industry began to suffer, especially in light of the high cost shares of cotton input in the low value added segments that dominate Pakistan’s textile industry. At the same time, the profitability of the sector probably suffered, as reflected in the steady decline in the market capitalization of the textile companies listed at Pakistan’s stock exchanges from 1994 or in the significant number of “sick” textile units (i.e., units against which banks hold nonperforming loans). Sluggish growth in recent years notwithstanding, the textile industry remains by far the most important industry in the country, reportedly employing 38 percent of the manufacturing labor force and accounting for 46 percent of all firms.

Pakistan: Cotton and Textile Developments, 1980–2000

Source: Data provided by the authorities; and IMF, International Financial Statistics.

See Hamid and others (1990) on trade and agricultural pricing policies in Pakistan until the late 1980s.

The geographical direction of exports

146. With the changing commodity composition of Pakistan’s exports, the geographical direction of exports shifted increasingly toward industrial countries’ markets over the last two decades. In the early 1980s, about 40 percent of Pakistan’s exports were directed to industrial countries. Over time, the share of exports to industrial countries rose to about60 percent. The United States became a particularly important destination of Pakistani exports, reflecting inter alia the increased production of sportswear and the demand for this product in the US market. Export shares to the major European trading partners, the United Kingdom and Germany, have been stable throughout the last two decades. The recent decline in the share of exports to Japan in total exports can be attributed to the stagnation in Japan during the 1990s. (The upper panel of Chart III-3 shows the shares of Pakistan’s exports to the five main export destinations.)

Chart III-3.Pakistan: Geographic Structure of Exports

Source: Data from Trade Analysis and Reporting System, United Nations.

1/ See text for details.

147. The proportion of Pakistan’s exports directed to developing countries generally decreased. Exports to Middle Eastern countries fell from an average of 10 percent between 1980–1985 to an average of about 8 percent during the period 1992–1998. Similarly, the share of exports to African markets fell from an average of 6 percent during the period 1980–1985 to 4 percent in 1992–1998, in part reflecting slower economic growth in this continent. There was a small increase in trading relations between Pakistan and other Asian countries, as measured by the share of Pakistan’ exports going to the Asian markets. This share increased from an average of 19 percent during 1980–85 to 22 percent during 1992–1998.

148. Pakistan’s export performance during the 1990s may not only be related to the commodity concentration of its exports but also to their geographical concentration. To capture the degree of geographic concentration in exports, a time series of Gini-Hirschman concentration indices was computed with the export shares by destination, based on export to 164 countries. The corresponding index values gradually rose from 0.21 in 1981 to 0.30 in 1998 (lower panel of Chart III-3). A general trend toward concentration was therefore not only a feature of the commodity composition of exports but also of the geographical destinations of exports. Data analysis confirms that the main reason for the increase is the striking rise in the share of exports to the United States, which increased from 5 percent in 1981 to 23 percent in 1998.

C. Comparing Pakistan’s Export Performance with Regional Competitors

149. In this section, Pakistan’s export performance is compared to that of major regional competitors. The comparison also includes the structure of exports and the trade regime to provide a cross-country perspective on the linkages between export performance and structure on the one hand and between performance and trade regime on the other.

Overall export performance

150. Pakistan’s exports evolved broadly in line with total world imports, as shown in Chart III-4. Accordingly, Pakistan’s share in world imports was remarkably stable during the last 20 years, ranging between a minimum of 0.12 percent in 1980 and a maximum of 0.18 percent in 1992. In 1999, the share was 0.14 percent. This would suggest that Pakistan’s export performance was not worse than that of the world on average. Compared to regional competitors, however, the performance was unimpressive, especially when compared to China and Thailand throughout the 1980s and 1990s or compared to Bangladesh, India, and Sri Lanka during the 1990s. All these countries succeeded in achieving sustained market share increases in total world imports.

Chart III-4.Paksitan: World Imports from Pakistan, and Comparator Countries

(Index of Imports in U.S. dollars; 1980 = 100)

Source: Data from Trade Analysis and Reporting System, United Nations; and staff calculations.

151. There are, of course, many factors that determine the export performance of a country. The diversification of exports by commodity groups and geographic destination are among these factors, at least in the short to medium term, as changes in the structure of exports are typically slow to occur.

The diversification of exports by commodity groups

152. To compare the diversification of exports of regional competitors, time series of Gini-Hirschman indices for each country were computed.73 The upper panel in Chart III-5 clearly shows that Pakistan’s high and increasing degree of concentration in exports by commodity groups was only surpassed by Bangladesh during the 1990s while that of Sri Lanka was close. Interestingly, the degree of commodity concentration also began to increase in Bangladesh from the late 1980s after having fallen during the early to mid-1980s. Except for Indonesia, other countries had stable and lower degrees of commodity concentration at the two-digit SITC level. Indonesia is the only country in which the commodity diversification of nonoil exports increased, as indicated by the fall in the Gini-Hirschman index value from about 0.45 in 1981 to 0.24 in 1998.

Chart III-5.Pakistan: Comparative Export Diversification

Source: Data from Trade Analysis and Reporting System, United Nations

1/ See text for details.

2/ Petroleum and Petroleum products category excluded from the calculations.

153. Comparing overall export performance and the degree of export diversification by commodity groups does not yield clear-cut conclusions. Good performers such as China or Thailand had a much more diversified export base, as was to be expected. However, Bangladesh did not benefit from a highly diversified export base but managed to position itself also among the better export performers from the mid-1980s to the late 1990s.

Geographic Diversification of Exports

154. Problems in the geographic concentration appear not to have been important in explaining Pakistan’s the disappointing trade performance in recent years. A comparison of Gini-Hirschman concentration indices for geographic distribution of exports of the same group of countries as above suggests that Pakistan’s exports are reasonably diversified in term of export destinations (lower panel in Chart III-5). The index values for Pakistan remained consistently below the panel mean value of 0.3 for all countries over time.

The position and structure of Pakistan’s textile exports

155. The analysis so far provides only limited evidence relating Pakistan’s less than impressive export performance in the 1990s to the general structure of exports. This raises the question to what extent Pakistan’s specialization on textiles contributed to the performance and how Pakistan’s textiles export performance fared compared with other countries.

156. Pakistan shares the feature of very high ratios of textile exports to total exports with Bangladesh, as the first panel in Chart III-6 shows. In Sri Lanka, textile exports also increased as a share of total export but from a lower base. In all other countries, shares of textile exports in total exports were lower and generally stagnant or even decreasing from the mid-1990s. Bangladesh’s favorable export performance during the 1990s as discussed above, which was achieved despite a high share of textiles, seems to repeat Pakistan’s experience during the 1980s and early 1990s. This suggests that the specialization on textiles may not be an obstacle to export growth over a decade or so, although the risks of the associated dependency on a narrow commodity base on balance of payments vulnerability and long-term growth prospects need to be kept in mind.

Chart III-6.Pakistan: Comparison of Textile Export Structure

Source: Data from Trade Analysis and Reporting System, United Nations.

157. As the contrast between the export performance in Pakistan and Bangladesh illustrates, a high share of textiles exports in total exports may not be sufficient to explain the prolonged recent stagnation in exports. The competitiveness of the textile sector is another important element. To gauge competitiveness problems in the textile sector, the values of textiles exports of the 6 comparator countries and Pakistan are compared in Chart III-7. Until 1991, Pakistan’s textile exports grew broadly in line with those of competitor countries. From 1992, however, Pakistan’s textile exports suffered from a noticeable slowdown compared to other countries, most notably China, Bangladesh, and Sri Lanka.

Chart III-7.Pakistan: Comparison of Textile Export Performance

(In billions of U.S. dollars)

Source: Data from Trade Analysis and Reporting System, United Nations; and staff calculations.

158. Without more detailed sectoral analysis, definitive conclusions about the reasons underlying the setback to the performance of Pakistan’s export performance cannot be drawn. Nevertheless, several factors relating to the textile sector appear to explain the weak overall export performance to some extent. The series of adverse shocks to the cotton crop from 1992, which reduced the supply of raw material and increased domestic cotton prices given the restrictions to cotton imports in place until recently, must be an important factor. One can also point out that Pakistan’s textile industry appears to have been less successful in upgrading to the higher value added segments of the industry, as the second panel in Chart III-6 shows. (Clothing generally is characterized by a higher value added content than textile fabrics such as yarn or cloth.) Unlike in other manufacturing sectors, upgrading is important for textile development and growth because of the Multi-Fiber Arrangement (MFA). The MFA regulates (and restricts) trade in textiles from developing to developed countries on the basis of quota allocations. Reportedly, Pakistan made full use of its quota allocations in the lower value added segments but not in the higher value added segments. Finally, the stagnant raw cotton prices during the 1990s, which largely determine the price for cotton-based textile fabrics in view of the low cost shares of other components, were another drag on the value of textile imports.

Trade regime

159. Despite significant progress toward a more liberal regime during the past 15 years (as described in Box III-1), Pakistan’s trade regime remains relatively restrictive by regional standards (Table III-2). As indicated by the trade restrictiveness index, only India’s trade regime remains more restrictive. Compared to other countries, the higher degree of restrictiveness is mainly due to higher average tariffs, as Pakistan has been lagging in catching up with other countries’ reductions in tariff rates and bands. In the reduction of non-tariff barriers, Pakistan appears to be at par with the comparator countries that have a more liberal overall trade regime. The impact of the trade regime on exports is difficult to evaluate because of other factors and measurement problems. Nevertheless, preliminary empirical analysis suggests that in Pakistan, a negative relationship held between exports (as a percent of GDP) and import taxes during 1973–99 (Box III-3). This indicates that the relatively higher tariff rates in Pakistan have been an obstacle to export performance, including through a tariff-induced anti-export bias in favor of production for domestic markets.74

Table III-2.Trade Regimes in Selected Asian Countries(2000 or latest unless otherwise noted)
Simple average



average tariff



Index of


Exports 1/Trade 1/
(In percent GDP)
Indonesia 2/6.83.34211.518.738.448.4
Sri Lanka11.47.45222.327.558.665.1
Sources: IMF, International Financial Statistics; and UN, Trade Analysis and Reporting System; and Fund staff estimates.

Period averages.

Nonoil merchandise exports.

Sources: IMF, International Financial Statistics; and UN, Trade Analysis and Reporting System; and Fund staff estimates.

Period averages.

Nonoil merchandise exports.

D. A Constant Market Shares Analysis of Pakistan’s Export Performance

160. This section applies the method of constant market share analysis to Pakistan’s export data. The method is particularly suited for analyzing the linkages between export performance and changes in the commodity composition of imports of trading partners. In the case of Pakistan, this question arises naturally given the high share of cotton-based textile exports, as the latter are often thought to be suffering from declining world demand compared to the demand for mixed fiber (man-made and natural) products.

The constant market shares method

161. Constant market share analysis is a method to study and assess export performance. The method decomposes the performance, which is measured on the basis of changes in the market share of a country’s exports in total imports of another market, into three components or effects.75 Increasing the market share over a period would mean success while a decrease would be interpreted as a problem. More specifically, the change in market share is decomposed as follows76:

Box III-3.The Impact of Trade Liberalization on Pakistan’s External Trade

Trade liberalization has enhanced the openness and external orientation of the Pakistan economy over the past two decades. Notwithstanding these gains, the ratio of external trade to GDP in Pakistan remains considerably lower than in most other economies in the region. This likely relates to the continued restrictiveness of Pakistan’s trade system, especially in relation to that of most other countries.

Cyclical factors and supply shocks (e.g. bad crops) complicate the quantitative assessment of the impact of trade liberalization on trade performance, since in the short term these factors can significantly reduce exports and indirectly curb imports as income levels fall. Moreover, as trade liberalization induced investment in the export sector may materialize with a lag, and since the liberalization in Pakistan has been undertaken gradually over almost two decades, it is even more difficult to isolate the relationship between trade performance and changes in the trade system.

Empirical estimates employing cointegration techniques indicate a significant long-run relationship between the trade regime and the external orientation of the Pakistan economy over the past 25 years. The long-run cointegration tests, which abstract from short-term effects of cyclical factors and supply shocks, consider the relationship between import taxes (defined as the ratio of customs duty collections to the value of imports), openness (defined as the ratio of exports and imports to GDP and, alternatively, as the ratio of exports to GDP), real GDP, and the time trend (which is intended to capture, collectively, the effect of the trend component of all missing variables in the equation).

The estimates indicate that, over the long run, a reduction in the import tax by 1 percentage point was associated with an increase of about 0.4 percent of GDP in the ratio of total trade to GDP and of 0.3 percent of GDP in the ratio of exports to GDP.1 The estimates also suggest that other factors—collectively captured by the time trend in the estimated equation—have tended to reduce Pakistan’s external orientation over the past two decades. While difficult to identify, these factors may have included the targeting of industrial policies in favor of the textiles sector and the decline in Pakistan’s access to external financing.

While the liberalization of the trade system has been associated with increased external trade, the performance of Pakistan’s exports over the past decade has been weaker than that in most other countries in the region (including Bangladesh, China, India, Indonesia, Korea, Malaysia, Philippines, Sri Lanka, and Thailand) and the ratio of external trade to GDP in Pakistan remains lower than in most of these countries. The value of Pakistan’s exports (in U.S. dollar terms) grew by an average of under 7 percent annually during 1990–99, while export growth in the other countries ranged from 9 percent to 16 percent a year. Pakistan’s trade to GDP ratio, which have raged 33 percent in the 1990s, was significantly lower than in all the other countries except India and Bangladesh. Although cyclical factors and supply shocks during the past decade may have weakened trade performance in Pakistan, it should be noted that Pakistan’s average import tariff rate remains considerably higher than in all the other countries except India.

By contrast, Subramanian et. al. (2000), in a study of trade liberalization in African countries, find that a 1 percentage point reduction in trade taxes was associated with an increase in the trade ratio of 0.7-1.1 percentage points.

where k stands for the exporting country and l for the importing country. It should be noted that all terms reflect aggregates over a range of commodities. The performance measure, the change in the total market share of exports from country k in total imports of country l (i.e., for all commodities i), is on the left hand side of the equation. The first right hand side term (ΔMSakl) captures the so-called market share effect, measuring the extent to which the exporting country’s market share for specific commodities has changed at the original commodity composition of the imports in the partner country. The second right-hand side term (ΔMSbkl) reflects the so-called commodity composition effect. This is the effect on which applications of the constant market share analysis have focused since Tyszynski (1951). It measures the effect of the changes in the commodity composition of country l’s imports on country k’s exports with the market shares in the base year. If the share of products on which country k is specialized increases between two periods, the effect is positive. The third right hand side term (ΔMSabkl), labeled as the commodity adaptation effect by Fagerberg and Sollie (1987), is a measure of the extent to which a country has succeeded in adapting the commodity composition of its exports to changes in the commodity composition of the importing country’s imports relative to other exporters. If this effect is zero, Fagerberg and Sollie (op. cit.) pointed out that the exporting country adapted to the commodity composition at the same rate as the average of all other countries.

162. In the current context, the commodity composition effect shows the extent to which the original commodity composition of exports (of the exporting country) was beneficial or problematic in view of the changes in the commodity composition of imports in the partner country. The commodity adaptation effect captures one dimension of competitiveness, namely that of competitiveness in adapting the commodity composition of exports to the need of the importing country. The market share effect can be interpreted as the residual that captures all other effects, including that of external demand by the partner country.

Empirical results

163. Constant market shares analysis was used to assess the change in Pakistani export market shares in four countries and regions from 1980 to 1997. The assessment is based on the decomposition of the market share changes over the entire sample period, which ranges from 1980 to 1997, and over sub-sample periods, 1980–1985, 1985–1990, 1990–1995, and 1995–1998.

164. The markets chosen were those of 15 members of the European Union (15 members), Japan, the United States and Asian Markets (excluding Japan). Pakistan’s export products are assumed to compete with exports from the rest of the world in these four markets. The analysis is based on export and import data that is disaggregated at the two-digit SITC code level. Only manufacturing sector data from the SITC categories 5–8 were used. This coverage in terms of markets and products covers about 60–65 percent of Pakistan’s exports in the 1990s and is expected to allow for a reasonable assessment of the overall performance of Pakistan’s export sector over the period.

165. The results of the application of the constant market shares analysis yields mixed results, depending on the recipient country and the time period (Table III-3).77 The salient features are as follows:

  • In the 15 EU countries, which constituted the largest market for Pakistan’s exports among the recipient countries, Pakistan was able to increase its market share throughout the sample period. Nevertheless, the increasing specialization turned out to have had adverse effects since the commodity composition effects turned out to be negative throughout all periods, as expected. As Pakistan’s specialization on textiles increased over time, the commodity adaptation effects were equally negative, except for the period 1980–85. In fact, the EU countries are a case in point for the argument of Pakistan being specialized in a shrinking market (relative to overall imports). The share of textile imports in total EU imports decreased throughout the period 1980–97 while the share of textiles in imports from Pakistan increased. Overall, the market share effects were the main factors behind the increase in Pakistan’s market share in EU imports.

  • In the United States, Pakistan was also able to increase its market share. The commodity composition effects were negative for the entire sample period and for 1985–90 but, unexpectedly, were positive for the periods 1980–85 and 1990–97. This reflects the increase in the share of textile imports in U.S. imports during these periods. Consequently, Pakistan was well positioned to benefit from this development with its increasing specialization in textiles and, as indicated by the positive commodity adaptation effects, was competitive in this sector. The commodity adaptation effects were positive in all periods, suggesting that Pakistan was competitive in the United States more generally. These results suggest that the case for the “textile pessimism” may not be as strong as often argued. As for the EU countries, the market share effects were the most important factors behind the increase in Pakistan’s market share in U.S. imports.

  • In Asian markets, the third largest import market in the sample after the EU countries and the United States, Pakistan lost market shares throughout 1980–97. The most important factor behind the losses was the commodity composition of Pakistan’s exports, as the negative commodity composition effects illustrate. One of the main factors behind this result is the decline in the imports of textile fabrics such as yarn or cloth by Asian countries. This is, of course, the segment in which Pakistan’s textile industry has specialized. The commodity adaptation effects are small, so that the loss in market shares appears not to be a reflection of problems in commodity composition competitiveness.

  • For exports to Japan, the results are varied. For the period of 1980–97, Pakistan was able to increase its market share somewhat. The commodity composition and commodity adaptation effects were negative, reflecting the decline in the share of imports of textile fabrics and leather products in Japan as well as some market share losses by Pakistan in the market for leather and leather products in Japan. With the increasing share of textile fabrics in total exports to Japan, Pakistan’s industry was not well positioned to take advantages of developments in Japan’s total imports.

Table III-3.Pakistan: Constant Market Shares Analysis(In percent of end-of-period total exports of Pakistan)
EEC Market1980-851985-19901990-19971980-1997
Market Share Effect3.17.95.617.1
Commodity Composition Effect-0.5-2.4-2.7-4.6
Commodity Adaptation Effect0.3-0.2-0.9-2.7
Change in Market Share2.
American Market1980-851985-19901990-19971980-1997
Market Share Effect0.
Commodity Composition Effect0.20.00.2-0.1
Commodity Adaptation Effect0.
Change in Market Share1.
Asian Market1980-851985-19901990-19971980-1997
Market Share Effect0.2-
Commodity Composition Effect0.1-1.8-3.6-4.0
Commodity Adaptation Effect-1.02.4-1.30.1
Change in Market Share-0.7-0.8-2.5-3.3
Japanese Market1980-851985-19901990-19971980-1997
Market Share Effect6.5-2.2-2.83.9
Commodity Composition Effect-0.2-3.2-1.6-1.9
Commodity Adaptation Effect-
Change in Market Share5.9-5.1-3.90.4
Source: Trade Analysis and Reporting System, United Nations; staff calculations.

The figures were derived by applying the percentage changes at the end-of-period imports of the partner region and normalizing them with the end-of-period total exports of Pakistan.

Source: Trade Analysis and Reporting System, United Nations; staff calculations.

The figures were derived by applying the percentage changes at the end-of-period imports of the partner region and normalizing them with the end-of-period total exports of Pakistan.

166. In general, the results show the benefits and risks associated with Pakistan’s increasing specialization on textiles. Except for the case of the United States, Pakistan export products have not evolved with the needs of the importing countries because they are becoming less important in the basket of import goods of the trading partners. Nevertheless, the generally negative commodity composition and commodity adaptation effects should not be overemphasized. As indicated by the generally positive market share effects, Pakistan has been able to benefit from the general growth in overall imports of the four trading partner covered in the analysis. This could be related partly to the general relocation of the production of textiles to developing countries. In the context of this relocation process, Pakistan succeeded in increasing its market shares in textile imports of partner countries despite the general tendency towards lower shares of textile products in total imports. The policy issues that emerge from the analysis are whether Pakistan benefited from this relocation process as much as it could have with optimal policies and why the textile industry was the only industry to have benefited from a more general relocation of industrial production to developing countries.

E. Conclusions and Policy Implications

167. The analysis of Pakistan’s merchandise export performance during the last two decades highlights the dependence of overall exports to the boom and bust cycle in the cotton sector. In the 1980s, the boom in the production of cotton helped to spur rapid growth in the downstream processing of cotton in the local textile industry. This, in turn, led to strong increases in textile exports, which gradually began to dominate Pakistan’s exports. The growing specialization on textiles helped Pakistan to benefit from the relocation of textiles production from industrial countries, even though the share of textile imports in total imports fell in some of these countries.

168. While the textile boom of the 1980s illustrated the benefits associated with specialization, the series of negative productivity shocks that hit the cotton sector from the mid-1990s demonstrated the risks associated with an undiversified export base. The shocks held back output growth in the textile industry as external trade in raw cotton was restricted and textile exports stagnated. The country-specific productivity shocks also appear to help in explaining why Pakistan’s export performance fell behind that of other regional competitors.

169. Pakistan’s experience with the boom and bust cycle in the cotton sector and its impact on the textile industry and exports during the last two decades has important policy implications for future trade reform and trade development:

  • While policies that targeted growth and trade in specific sectors such as the textiles industry were initially successful in boosting exports, they also contributed to increasing the vulnerability to sector-specific shocks. The initial success of sector-specific policies may also explain why policies aimed at fostering broad-based growth in the tradables sector, including exchange rate policy, may have played less of role. For the future, measures to strengthen the diversification of the export commodity base will be essential for balanced growth in the tradables sector and the economy.

  • Pakistan’s experience also illustrates how targeted trade restrictions can backfire and hurt long-term development. In Pakistan, relevant restrictions included the general ban on imports of raw cotton78 (removed in 1999/2000), the government monopoly on the exports of raw cotton (removed in the mid-1990s), and the ban on many semi-manufactured textile inputs such as cloth. With the adverse shocks to cotton output, economic activity in the textile sector was handicapped by the restrictions on imports of raw cotton. Further progress in trade liberalization, to which Pakistan has committed, will be needed to remove the anti-export bias in today’s trade regime and thereby increase the medium-term growth potential.

  • Industrial policies appear to have focused too little on promoting public goods such as standards—for example, in setting standards for the quality of lint cotton—and providing essential public services including in the area of infrastructure services. Instead, policies appear to have been focused too much on providing tax and other incentives, which may have fostered overinvestment in some segments of the textile industry. This overinvestment, which resulted in a debt overhang in the sector, may have slowed the upgrading process in the industry in the 1990s. In the future, government efforts should focus on creating the enabling environment for tradables sector growth through the increased provision of the necessary physical and institutional infrastructure (including norms and standards).

  • The combination of trade restrictions and an interventionist industrial policy targeting specific sectors through incentives may have reduced competitive pressures on the textile industry and provided incentives for rent-seeking activities. As a result, market pressures for upgrading and modernization in the textile industries were mitigated despite the evolving structure of the world textile industry and the world demand for textiles. Unfortunately, this has hampered long-term growth prospects of the industry, and as widely recognized in Pakistan, a lot of catching-up by the industry is needed. This is yet another reason for more trade liberalization and for reorienting efforts to generate the enabling environment for broad-based growth in tradables to reduce the risks of recurring boom and bust cycle elsewhere.

170. Looking forward, policies aimed at trade reform and strengthening the tradables sector in general will also help in addressing the challenges arising from the on-going evolution in the international trade regime. The implementation of agreements under the Uruguay Round accord offers a variety of opportunities for Pakistan to expand its export base and diversify its export markets. The most promising opportunities are likely to come from multilateral trade liberalization affecting the products and sectors, in which Pakistan enjoys a comparative advantage. Of particular relevance to Pakistan is the phasing out of the MFA. As Pakistan’s textile and clothing exports are largely directed to quota countries, and since quota utilization rates in many segments are very high, Pakistan should benefit from the eventual removal of the MFA in 2005. Ingno and Winter (1995) estimate that Pakistan may gain more than US$500 million (1992 prices) or a permanent increase in export levels of about 6 percent. Reaping the benefits from the phasing out of the MFA, however, will neither come quickly nor without overcoming challenges. Pakistan will face fierce competition, particularly from Eastern Europe countries, which benefit from skilled and relatively low wage labor and proximity to EU market.

APPENDIX III-I Constant Market Share Analysis: The Case of Several Commodities and One Importing Country79

The following symbols are used in decomposing the change in market shares into market composition, commodity composition and market adaptation effects.

nnumber of commodities;
0, tthese are the initial year and the final year of comparison, respectively;
Yiklcountry k’s exports of commodity i to country l;
Bilcountry l’s imports of commodity i;
MSklmarket share of country k (macro share of country k) in country l’s imports;
aklmarket shares, by commodity, of country k (micro shares of country k) in country l’s imports; row vector of dimension n; akl = (a1kl,………….., ankl), where
blcommodity shares of country l’s imports; column vector of dimension n; bl=(bil,,bnl), where

The macro share of country k (MSkl) may be written as the inner product of the vector of its micro shares (akl) and the vector of commodity shares of country l’s imports (bl);

The change in MSkl between time 0 and time t is given by (5):

Equation (5) can be rewritten as:


The first of these terms (ΔMSakl) is the so-called market share effect, and the second term (ΔMSbkl) reflects the commodity composition effect. The third (residual) term (ΔMSabkl) is the inner product of a vector of changes in micro shares and a vector of changes in commodity shares and was labeled as the commodity adaptation effect by Fagerberg and Sollie

Prepared by Adedeji Olumuyiwa (INS) and Thomas Helbling (MED), with input from Aasim Husain (MED), Marcio Ronci (PDR), and Farhan Hameed (MED).

The reference is to fiscal years. Pakistan’s fiscal years begin on July 1 and end on June 30. For example,1981 would refer to the fiscal year 1980/81.

Trend growth rates in Chart III-1 are the growth rates of the trend component in the log-level of the export series, which was derived with a bandpass filter (see Baxter and King, 1999). The last two years of trend growth are interpolated. As in the case of the Hodrick-Prescott filter, the end-point estimates provided by the bandpass filter are only indicative.

The Gini-Hirschman trade concentration index is defined as G=Σi(xi/x)2, where xi, is the value of a country’s exports in commodity category i, while x is the value of total trade (see Hirschman, 1964).

It should be noted that the finding of a rising concentration would emerge even if exports of raw cotton were included in the textiles aggregates.

Trade restrictions also included bans on the imports of semi-manufactures such as cloth or products made of artificial fibers.

The index calculations are again based on data at the two-digit SITC code level.

See also Khan (1998).

See the following studies on the evolution of constant market analyses as a technique for assessing export performance: Tyszynski (1951); Richardson (1971); and Fagerberg and Sollie (1987).

See Appendix III-I for the mathematical underpinning of these effects.

The constant market share analysis is based on imports of the partner country, e.g., the United States. The dimensions of the three effects and the performance measures are percentage point changes in terms of the imports of the partner countries. However, as Pakistan’s exports to any of the four partner countries covered in Table III-3 are less than 1 percent of total imports, the magnitude of the market share changes and effects is small (0.1 percentage point or less). To facilitate the interpretation of Table III-3, the results were renormalized to compare them to Pakistan’s export. The this end, the percentage changes were first converted into U.S. dollar values by multiplying them with the end-of-period import values of the partner country (e.g., 1997 total U.S. imports) and then normalized by the end-of-period total exports of Pakistan. With this normalization, the magnitude of each effect over the entire period does not equal the sum of each effect over the subsample periods.

The state-owned Trading Corporation of Pakistan (TCP) did, however, occasionally import small quantities of cotton.

This exposition follows Fagerberg and Sollie (1987) closely.

Other Resources Citing This Publication