## 1. India's Annual Exports, 1980–2001

## Source details

**Canonical URL:** [1. India's Annual Exports, 1980–2001](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04173.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04173.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04173.pdf.json)

---

### Introduction and research questions
- India liberalized in 1991 amid a balance of payments crisis; total exports of goods and services rose from about US$18 billion in 1991 to about $51 billion in 2002.
- Research goal: understand determinants of the large export increase by focusing on microeconomic (firm-level) foundations of exporting success.
- Links literature strands: (a) trade and FDI liberalization → firm productivity; (b) determinants of firm export entry (productivity, sunk costs); (c) FDI as export catalyst via technology/information spillovers and competition.
- Two principal channels for multinational (MNC) effects on domestic exporters:
  - Demonstration/information and technological spillovers when MNCs export (local firms learn foreign-market distribution/marketing).
  - Increased domestic competition from MNC entry, inducing incumbents to innovate and adopt leaner production techniques.
- Empirical focus: test whether MNCs acted as export catalysts through demonstration effects or competitive-pressure channel using firm-level data from India’s post-1991 liberalization.

### Key empirical questions and approach
- Decompose export growth into four contributors: entrants (new firms), exiting firms, reallocation to more export-intensive firms, and increased export-orientation of pre-existing firms.
- Estimate dynamic discrete choice models for firm exporting status to test for sunk-cost hysteresis (dependence on prior exporting experience).
- Test effects of firm characteristics (productivity, profitability, size, capital intensity), ownership (multinational, public sector), and FDI-related variables (MNC export concentration, local export concentration, MNC market share) on export probability.
- Use nonstructural (reduced-form) binary choice framework; address dynamic panel issues with fixed effects, first-differences IV, and Arellano–Bond GMM.

### Main empirical findings
- Decomposition result (1991–2000): surviving incumbent firms drive most of the increase in aggregate export intensity.
  - Export Intensity Gains for Survivors: 46.5 percent
  - Reallocation among Surviving Firms: 24.8 percent
  - Contribution of Entering Firms: 2.3 percent
  - Contribution of Exiting Firms: 26.4 percent
  - Interpretation: surviving firms’ internal increases and reallocation account for over 73 percent of change.
- Sunk costs and persistence:
  - Lagged exporting status is a strong predictor of current exporting status; dynamic specifications find substantial sunk costs to exporting.
- Firm characteristics:
  - More productive firms are more likely to export; productivity measure used is gross value added over total wage bill.
  - Sales (size) and profitability (profits/assets; profit margins over sales) are positively associated with export probability.
- Ownership effects:
  - Multinational dummy is positive and significant—being an MNC increases the likelihood of exporting (about 3.5 percent in one specification).
  - Public sector firms are less likely to export—being a public sector manufacturing firm reduces export probability by more than 5 percent in results reported.
  - Distinctions among private ownership categories (top 50 firms, large business houses, foreign business houses, NRIs) were generally not significant.
- FDI-related channels:
  - No evidence found for localized informational or technological spillovers from MNC export concentration or local export concentration (lagged MNC export concentration and lagged local export concentration in state-industry are insignificant).
  - MNC market share in the industry is positive and significant—consistent with competitive-pressure channel: larger MNC market share increases likelihood that a firm will export.
- Location:
  - Coastal state dummy positive and significant in levels specification.
- Robustness:
  - Levels OLS without firm fixed effects shows lagged export status accounts for approximately 76 percent of likelihood of exporting (note potential overestimation due to unobserved heterogeneity).
  - Fixed effects and Arellano–Bond procedures applied to address bias and unobserved heterogeneity; preferred specification corrects downward bias on lagged dependent variable by estimating and adding back Nickell (1981)-type bias for moderately large T.

### Illustrative case: Indian auto sector
- Pre-1993: highly protected; dominant domestic firms (Hindustan Motors, PAL) operated well below global best-practice productivity.
- Gradual and then full opening to FDI: Suzuki entered in 1983 via joint venture with Maruti; sector fully opened in 1993 leading to broad MNC entry.
- Investment and market entry:
  - Roughly $1.6 billion invested after opening.
  - Twelve new players entered since 1993.
- Industry outcomes:
  - Real prices fell on average by 2–6 percent between 1998 and 2001.
  - Sector-level profitability declined by about 25 percent in 2001–02, largely due to real price declines.
  - Lowest productivity manufacturers exited; Hindustan Motors market share fell from 100 percent (pre-1983) to about 3 percent by 2003.
  - Maruti (joint venture) productivity grew at a compound annual growth rate of 10 percent since 1993 (productivity measured as cars produced per employee).
  - Industry productivity increased fivefold since 1983; India now produces 13 times more cars than in 1983; exports grew from zero to roughly 10 percent of production.
  - Components industry: annual export growth in excess of 40 percent.
- Interpretation: FDI had a strong positive impact via competitive dynamics forcing incumbents to reform or exit and enabling highly productive firms to succeed as exporters.

### Data sources and sample characteristics
- Aggregate export analysis: UN Comtrade database (overall and by-industry exports).
- Firm-level analysis: Prowess dataset (compiled by CMIE) — balance sheet information on over 6000 mostly listed firms, accounting for more than 70 percent of organized industrial sector activity.
- Stylized export trends:
  - Exports started rising in 1986 but accelerated post-1991; three distinct periods in the 1990s: rapid acceleration from 1991, slowdown 1995–1999, pickup from 1999.
  - India’s share of world exports increased from 0.5 to 0.8 percent in the period 1991–2002.
- Sample summary (selected exact figures from Table 1):
  - Total # of firms: 2065 (1991) … 5558 (2001)
  - # of firms exporting: 1090 (1991) … 2516 (2001)
  - % of firms exporting: 52.8 (1991) … 45.3 (2001)
  - # of MNCs: 209 (1991) … 403 (2001)
  - % of MNCs exporting: 68.9 (1991) … 64.0 (2001)
  - # of Govt. Firms: 227 (1991) … 293 (2001)
  - % of Govt. Firms exporting: 44.5 (1991) … 34.5 (2001)
  - # of Entering Firms (selected years): 350 (1993) … 540 (2001)
  - % of Entering firms exporting (selected): 27.1 (1993) … 38.2 (2001)

### Empirical methodology — concise points
- Theoretical basis: firm exports if expected discounted profits net of entry costs are positive; entry costs (sunk) create hysteresis and intertemporal link in export decision.
- Reduced-form binary choice model estimated across specifications: levels OLS (no firm effects), levels with firm fixed effects (bias-corrected for lagged dependent), first-differences with instruments, Arellano–Bond GMM.
- Address endogeneity of lagged exporting status via (a) fixed effects bias correction following Nickell (1981) for moderately large T, and (b) Arellano–Bond GMM using lagged levels as instruments.
- FDI measures:
  - MNC export concentration in state-industry: state-industry MNC exports as share of national industry exports relative to state share of national manufacturing exports.
  - Local export concentration analogously defined using overall state-industry exports.
  - MNC market share in industry used as proxy for competitive pressure.

### Estimation results and dynamic bias adjustment
- The extent of the Nickell (1981) bias is computed as 0.11.
- Adding the bias to the coefficient on the lagged dependent variable gives an unbiased and consistent estimate of 0.45.
- Arellano-Bond (specification (3)) finds a coefficient of 0.43, suggesting robustness across estimation methods.
- Selected coefficients and standard errors from Table 4 (Fixed Effects and First Difference-IV Model of Export Participation):
  - Exported Last Year: 0.34**; 0.34**; 0.43**  
    - Standard errors: 0.005; 0.005; 0.01
  - Productivity: 0.03**; 0.04**; 0.007  
    - Standard errors: 0.01; 0.01; 0.01
  - Profitability: 0.02**; 0.02**; -0.004  
    - Standard errors: 0.008; 0.008; 0.01
  - MNC market share in industry: 0.15**; 0.15**; -0.06  
    - Standard errors: 0.06; 0.06; 0.12
  - Local Export Conc. In State-Industry: 0.002  
    - Standard error: 0.01
  - MNC Export Conc. in State-Industry: 0.02; 0.003  
    - Standard errors: 0.20; 0.03
  - Year Dummies: Yes; Yes; Yes
  - No. of Observations: 35493; 35493; 29085
  - Overall R-squared: 0.59; 0.59
  - F-statistic: 288.4; 288.4
  - Wald Chi-square: 1857
- Note: Binary Dependent Variable: Y=1 if exporter; ** Significant at 1 percent level; * Significant at 5 percent level. Profitability measured as profits/assets; Productivity measured as gross value added/wage bill.

### Specification differences, instruments, and robustness
- Fixed effects regressions:
  - Productivity remains significant and positive after controlling for firm fixed effects.
  - Higher profits on assets increase the probability of exporting.
  - Fixed effects regressions drop time-invariant variables such as ownership group, age, and reforms dummy.
- Arellano-Bond (FD-IV) specification:
  - None of the firm characteristics are significant in the FD-IV specification.
  - Possible reasons: firm characteristics may be level effects and insignificant in differences; weak instruments complicate FD-IV estimates.

### FDI, spillovers, and market-structure effects
- Fixed effects specifications: spillover variables (MNC export concentration, local export concentration) are insignificant — little evidence of localized spillovers from MNC exporting activity stimulating firm exports.
- MNC market share in the industry is positive and highly significant in fixed effects regressions — consistent with competitive-pressure channel.
- Multinational activity after 1991 largely horizontal and aimed at the domestic market; demonstration effects for local firms are small.
- More disaggregated industry checks produced positive effects of MNC market share on export probability but are not reported due to non-random selection bias.

### Interpretation of dynamics and market outcomes
- Large sunk costs to exporting: incumbent firms better able to take advantage of trade reforms and improved exporting conditions.
- Entry of MNCs increased industry competitiveness, reduced profit margins, and increased the probability a firm will export.
- More productive firms more likely to export; multinationals have higher export probability; government firms less likely to export.
- Coastal state firms more likely to export, consistent with access-to-coast advantages.

### Caveats and data limitations
- Sample excludes small, unregistered firms; potential sample selection if small new export-oriented firms differ from larger registered firms.
- Possible unobserved small local firms learning to export from multinationals but not captured in the sample.
- Geographical spillover measurement uses headquarters location as firm location; plant-level data would be required to correct for multi-plant dispersion.
- Endogeneity concerns that MNC market share could be higher in more open industries are mitigated by evidence that MNC investment in India is primarily for the local market.

### Evolution of Indian export performance and policy implications
- 1991 reforms and devaluation provided initial impetus; trade-weighted tariff rates fell from 87 percent in 1991 to 25 percent by 1997; nontariff barriers were also reduced.
- Most increase in export intensity during the reform period was driven by existing firms becoming more export-oriented rather than by large-scale entry of new export-oriented firms.
- After 1999, export growth resumed, driven largely by services exports (mainly IT-related products) and manufactured goods; increased productivity of Indian firms played a large role.
- Contrast with China: FDI in China was largely export-oriented and catalyzed entry of new export-oriented SMEs; in India, FDI acted as a competitive spur forcing incumbents to innovate.

- Policy-relevant recommendations:
  - Focus on incumbent firm capabilities: much export growth arises from existing firms becoming more export intensive.
  - Policies facilitating competition (including through FDI entry) can raise export performance via productivity improvements induced by competitive pressure.
  - Direct attempts to rely on localized spillovers from MNC export activity may be less effective; evidence points to competition (market share effects) rather than spatial export-concentration spillovers as the FDI-to-export channel.
  - Address sunk entry costs to exporting (information, distribution, regulation compliance, credit): prior exporting experience strongly raises future export probability.
  - Policies should emphasize easing bottlenecks (information, infrastructure, bureaucratic procedures) rather than providing direct subsidies based on export value.
  - Evidence argues against ‘infant industry’ protection; competition has spurred innovation leading to export growth.

*Source: _wp04173 - 1.  India's Annual Exports, 1980–2001 (PDF chapter/section).*

### 1.  India's Annual Exports, 1980–2001...................................................................................

### 1. India's Annual Exports, 1980–2001

### Sectoral Growth Rates in the 1990s
- Section present on page 15.
- Focus: sectoral growth rates during the 1990s (section title preserved exactly: "Sectoral Growth Rates in the 1990s").

### Largest Exporting Industries
- Section present on page 16.
- Focus: identification and discussion of the largest exporting industries (section title preserved exactly: "Largest Exporting Industries").

### Tables and Empirical Components
- Tables listed in the content unit (with exact titles and page references where provided):
  - Table 1. Sample Characteristics — page 16
  - Table 2. Decomposition of Growth in Aggregate Export Intensity, 1991–2000 — page 18
  - Table 3. Firm Characteristics and the Decision to Export — page 20
  - Table 4. Fixed Effects and First Difference-IV Model of Export Participation — page 22

*Source: _wp04173 - 1.  India's Annual Exports, 1980–2001 (PDF chapter/section).*

### References..............................................................................................................

### References

### I. Introduction — context and research questions
- India liberalized in 1991 amid a balance of payments crisis; total exports of goods and services rose from about US$18 billion in 1991 to about $51 billion in 2002.
- Research goal: understand determinants of large export increase by focusing on microeconomic (firm-level) foundations of exporting success.
- Links literature strands: (a) trade and FDI liberalization → firm productivity; (b) determinants of firm export entry (productivity, sunk costs); (c) FDI as export catalyst via technology/information spillovers and competition.
- Two principal channels for multinational (MNC) effects on domestic exporters:
  - Demonstration/information and technological spillovers when MNCs export (local firms learn foreign-market distribution/marketing).
  - Increased domestic competition from MNC entry, inducing incumbents to innovate and adopt leaner production techniques.
- Empirical focus: test whether MNCs acted as export catalysts through demonstration effects or competitive-pressure channel using firm-level data from India’s post-1991 liberalization.

### Key empirical questions and approach
- Decompose export growth into four contributors: entrants (new firms), exiting firms, reallocation to more export-intensive firms, and increased export-orientation of pre-existing firms.
- Estimate dynamic discrete choice models for firm exporting status to test for sunk-cost hysteresis (dependence on prior exporting experience).
- Test effects of firm characteristics (productivity, profitability, size, capital intensity), ownership (multinational, public sector), and FDI-related variables (MNC export concentration, local export concentration, MNC market share) on export probability.
- Use nonstructural (reduced-form) binary choice framework; address dynamic panel issues with fixed effects, first-differences IV, and Arellano–Bond GMM.

### Main empirical findings — summary
- Decomposition result (1991–2000): surviving incumbent firms drive most of the increase in aggregate export intensity.
  - Export Intensity Gains for Survivors: 46.5 percent
  - Reallocation among Surviving Firms: 24.8 percent
  - Contribution of Entering Firms: 2.3 percent
  - Contribution of Exiting Firms: 26.4 percent
  - Interpretation: surviving firms’ internal increases and reallocation account for over 73 percent of change.
- Sunk costs and persistence:
  - Lagged exporting status is a strong predictor of current exporting status; dynamic specifications find substantial sunk costs to exporting.
- Firm characteristics:
  - More productive firms are more likely to export; productivity measure used is gross value added over total wage bill.
  - Sales (size) and profitability (profits/assets; profit margins over sales) are positively associated with export probability.
- Ownership effects:
  - Multinational dummy is positive and significant—being an MNC increases the likelihood of exporting (about 3.5 percent in one specification).
  - Public sector firms are less likely to export—being a public sector manufacturing firm reduces export probability by more than 5 percent in results reported.
  - Distinctions among private ownership categories (top 50 firms, large business houses, foreign business houses, NRIs) were generally not significant.
- FDI-related channels:
  - No evidence found for localized informational or technological spillovers from MNC export concentration or local export concentration (lagged MNC export concentration and lagged local export concentration in state-industry are insignificant).
  - MNC market share in the industry is positive and significant—consistent with competitive-pressure channel: larger MNC market share increases likelihood that a firm will export.
- Location:
  - Coastal state dummy positive and significant in levels specification.
- Robustness:
  - Results reported from levels OLS without firm fixed effects show lagged export status accounts for approximately 76 percent of likelihood of exporting (note potential overestimation due to unobserved heterogeneity).
  - Fixed effects and Arellano–Bond procedures applied to address bias and unobserved heterogeneity; preferred specification corrects downward bias on lagged dependent variable by estimating and adding back Nickell (1981)-type bias for moderately large T.

### Illustrative case: Indian auto sector
- Pre-1993: highly protected, dominant domestic firms (Hindustan Motors, PAL) operated well below global best-practice productivity.
- Gradual and then full opening to FDI: Suzuki entered in 1983 via joint venture with Maruti; sector fully opened in 1993 leading to broad MNC entry.
- Investment and market entry:
  - Roughly $1.6 billion invested after opening.
  - Twelve new players entered since 1993.
- Industry outcomes:
  - Real prices fell on average by 2–6 percent between 1998 and 2001.
  - Sector-level profitability declined by about 25 percent in 2001–02, largely due to real price declines.
  - Lowest productivity manufacturers exited; Hindustan Motors market share fell from 100 percent (pre-1983) to about 3 percent by 2003.
  - Maruti (joint venture) productivity grew at a compound annual growth rate of 10 percent since 1993 (productivity measured as cars produced per employee).
  - Industry productivity increased fivefold since 1983; India now produces 13 times more cars than in 1983; exports grew from zero to roughly 10 percent of production.
  - Components industry: annual export growth in excess of 40 percent.
- Interpretation: FDI had a strong positive impact via competitive dynamics forcing incumbents to reform or exit and enabling highly productive firms to succeed as exporters.

### Data sources and sample characteristics
- Aggregate export analysis: UN Comtrade database (overall and by-industry exports).
- Firm-level analysis: Prowess dataset (compiled by CMIE) — balance sheet information on over 6000 mostly listed firms, accounting for more than 70 percent of organized industrial sector activity.
- Stylized export trends:
  - Exports started rising in 1986 but accelerated post-1991; three distinct periods in the 1990s: rapid acceleration from 1991, slowdown 1995–1999, pickup from 1999.
  - India’s share of world exports increased from 0.5 to 0.8 percent in the period 1991–2002.
- Sample summary (selected exact figures from Table 1):
  - Total # of firms: 2065 (1991) … 5558 (2001) [table shows year-by-year counts]
  - # of firms exporting: 1090 (1991) … 2516 (2001)
  - % of firms exporting: 52.8 (1991) … 45.3 (2001)
  - # of MNCs: 209 (1991) … 403 (2001)
  - % of MNCs exporting: 68.9 (1991) … 64.0 (2001)
  - # of Govt. Firms: 227 (1991) … 293 (2001)
  - % of Govt. Firms exporting: 44.5 (1991) … 34.5 (2001)
  - # of Entering Firms (selected years shown in table): ... 350 (1993) … 540 (2001)
  - % of Entering firms exporting (selected): ... 27.1 (1993) … 38.2 (2001)

### Empirical methodology — concise points
- Theoretical basis: firm exports if expected discounted profits net of entry costs are positive; entry costs (sunk) create hysteresis and intertemporal link in export decision.
- Reduced-form binary choice model estimated:
  - Key regression forms: levels OLS (no firm effects), levels with firm fixed effects (bias-corrected for lagged dependent), first-differences with instruments, Arellano–Bond GMM.
  - Address endogeneity of lagged exporting status due to unobserved firm heterogeneity by (a) fixed effects bias correction following Nickell (1981) approach for moderately large T, and (b) Arellano–Bond GMM using lagged levels as instruments in first differences.
- FDI measures constructed to distinguish channels:
  - MNC export concentration in state-industry: state-industry MNC exports as share of national industry exports relative to state share of national manufacturing exports.
  - Local export concentration analogously defined using overall state-industry exports.
  - MNC market share in industry used as proxy for competitive pressure.

### Policy-relevant implications (from findings)
- Export promotion and productivity policies should focus on incumbent firm capabilities: much export growth arises from existing firms becoming more export intensive rather than from entrants.
- Policies facilitating competition (including through FDI entry) can raise export performance via productivity improvements induced by competitive pressure.
- Direct attempts to rely on localized spillovers from MNC export activity may be less effective; empirical evidence points to competition (market share effects) rather than spatial export-concentration spillovers as the FDI-to-export channel.
- Addressing sunk entry costs to exporting (information, distribution, regulation compliance, credit) remains important because prior exporting experience strongly raises future export probability.

*Italic: Source: _wp04173 - References..............................................................................................................*

### Section IV earlier. The extent of the bias is computed using the Nickell (1981) method, and is

### _wp04173 - Section IV earlier. The extent of the bias is computed using the Nickell (1981) method, and is

### Estimation results and dynamic bias adjustment
- The extent of the Nickell (1981) bias is computed as 0.11.
- Adding the bias to the coefficient on the lagged dependent variable gives an unbiased and consistent estimate of 0.45.
- Arellano-Bond (specification (3)) finds a coefficient of 0.43, suggesting robustness across estimation methods.
- Table 4 (Fixed Effects and First Difference-IV Model of Export Participation) — selected coefficients and standard errors (standard errors below coefficients):
  - Exported Last Year: 0.34**; 0.34**; 0.43**  
    - Standard errors: 0.005; 0.005; 0.01
  - Productivity: 0.03**; 0.04**; 0.007  
    - Standard errors: 0.01; 0.01; 0.01
  - Profitability: 0.02**; 0.02**; -0.004  
    - Standard errors: 0.008; 0.008; 0.01
  - MNC market share in industry: 0.15**; 0.15**; -0.06  
    - Standard errors: 0.06; 0.06; 0.12
  - Local Export Conc. In State-Industry: 0.002  
    - Standard error: 0.01
  - MNC Export Conc. in State-Industry: 0.02; 0.003  
    - Standard errors: 0.20; 0.03
  - Year Dummies: Yes; Yes; Yes
  - No. of Observations: 35493; 35493; 29085
  - Overall R-squared: 0.59; 0.59
  - F-statistic: 288.4; 288.4
  - Wald Chi-square: 1857
- Note: Binary Dependent Variable: Y=1 if exporter; ** Significant at 1 percent level; * Significant at 5 percent level. Profitability measured as profits/assets; Productivity measured as gross value added/wage bill.

### Firm characteristics, specification differences, and instrument concerns
- Fixed effects regressions:
  - Productivity is significant and positive even after controlling for firm fixed effects.
  - Higher profits on assets increase the probability of exporting.
  - Fixed effects regressions drop time-invariant variables such as ownership group, age, and reforms dummy.
- Arellano-Bond (FD-IV) specification:
  - None of the firm characteristics are significant in the FD-IV specification.
  - Possible reasons: firm characteristics such as productivity, returns on assets, and profit margins are level effects and may no longer be significant in the differences specification.
  - Weak instruments complicate the FD-IV version, leading to insignificant coefficients.

### FDI, spillovers, and market structure effects
- In fixed effects specifications in Table 4, none of the spillover variables are significant, suggesting little evidence that spillovers from general exporting activity or multinational activity have had a large role in stimulating firms to export.
- The variable for MNC market share in the industry is positive and highly significant in fixed effects regressions, consistent with the hypothesis that foreign presence increases competitive pressure on local firms and indirectly increases their probability of exporting.
- Multinational activity after 1991 has largely been horizontal and aimed at the domestic market; hence, demonstration effects for local firms are small and MNC export concentration does not show positive spillovers on firm exports.
- Robustness check at more disaggregated industry level (four to five digit ITC classification) produced significant and positive effects of MNC market share on probability of exports, but these results are not reported due to non-random selection bias.

### Interpretation of dynamics and market outcomes
- Large sunk costs to exporting:
  - Incumbent firms have been better able to take advantage of trade reforms and improved exporting conditions.
  - Survivors are more likely to be better performers and have driven increased exporting activity.
- Entry of MNCs increased industry competitiveness, reduced profit margins, and increased the probability a firm will export.
- Firms that are more productive are more likely to export, consistent with foreign participation raising productivity and hence export propensity.
- Multinationals have a higher probability of exporting; government firms have a lower tendency to export — implying a declining public sector share of GDP should encourage higher export activity.
- Firms in coastal states are more likely to export, consistent with access-to-coast advantages seen in other contexts.

### Caveats and data limitations
- Sample excludes small, unregistered firms; potential sample selection if small new export-oriented firms differ from larger registered firms.
- Possible unobserved small local firms learning to export from multinationals but not captured; this would require that small unregistered firms harness externalities better than larger registered firms.
- Geographical spillover measurement uses headquarters location as firm location; plant-level location data would be required to correct for multi-plant dispersion.
- Endogeneity concerns that MNC market share could be higher in more open industries are mitigated by evidence that MNC investment in India is primarily for the local market (low share of exports from MNC total production and industry types).

### Discussion: evolution of Indian export performance (1990s onward)
- 1991 reforms and devaluation provided initial impetus for exports; trade-weighted tariff rates fell from 87 percent in 1991 to 25 percent by 1997. Nontariff barriers such as licensing requirements were also reduced.
- Most increase in export intensity during the reform period was driven by existing firms becoming more export-oriented rather than large-scale entry of new export-oriented firms.
- By 1995, trade reforms slowed, new entrants increased competitive pressure and squeezed profit margins; subsequent shakeout and efficiency responses followed.
- After 1999, export growth resumed, driven largely by services exports (mainly IT-related products) and manufactured goods; increased productivity of Indian firms played a large role.
- Contrast with China: FDI in China was largely export-oriented and catalyzed entry of new export-oriented SMEs; in India, FDI acted as competitive spur forcing incumbents to innovate and become competitive.

### Policy implications and recommendations
- Policy should distinguish between:
  - Exporters attempting to export more (incumbent firms increasing export intensity).
  - New firms attempting to enter export markets.
- If entry costs are important, policies should aim to ease bottlenecks rather than provide direct subsidies based on export value:
  - Provide information about potential markets.
  - Develop exporting infrastructure.
  - Eliminate bureaucratic hurdles to export.
- Evidence argues against ‘infant industry’ protection; competition has spurred innovation leading to export growth.
- Policymakers should focus on fostering competition and innovation (including liberalization of FDI rules) so incumbent firms near the technology frontier can become more productive and increase exports, rather than relying on protection and export incentives.

*Italic: IMF Working Paper section content as provided in the source content unit.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04173.pdf_