## _wp07130

## Source details

**Canonical URL:** [_wp07130](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07130.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07130.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07130.pdf.json)

---

### I. Introduction — scope and central questions
- Focus:
  - Examines relative growth performance across Asia at aggregate and sectoral levels.
  - Analyzes determinants of resource shifts from low- to high-productivity sectors.
- Key questions:
  - Whether development paths of Japan and the NIEs are being followed by ASEAN-4, China, India, and newly emerging economies such as Vietnam.
  - Whether systematic differences exist between East Asia and the rest of Asia, or between Asia and other regions.
  - Sources of growth differences within Asia and relative to other regions, and the role of policies.
  - How Asia’s high growth rates can be sustained and whether post-Asian-crisis reforms have had detectable impacts on growth and productivity.

### II. Asia’s economic success — stylized facts and comparative patterns
- Long-run performance:
  - Asia’s real income per capita rose sevenfold between 1950 and 2005.
  - Significant reduction in the gap relative to the United States; contrast with Latin America and other developing economies that failed to catch up.
- Within-region heterogeneity:
  - Per capita income range example: over $33,000 in Singapore to $2,000 in Bangladesh.
  - Takeoff dating used: Japan 1955; NIEs 1967; ASEAN-4 1973; China 1979; India 1982; Other Asia 1990.
- Trade, education, demographics:
  - Asia’s share of world trade more than doubled during 1970–2005.
  - Rapid increase in educational attainment over the last half century.
  - Declining dependency ratios supported growth; in Indonesia, Korea, Taiwan Province of China the demographic impact reached more than 1 percentage point per year.
  - NIEs and China likely to see rising dependency ratios within the next five years; India’s demographic transition started relatively recently.
- Policy and institutional environment:
  - Fast developers: broadly stable macroeconomic environments and contained inflation (exceptions: oil shocks and the 1997 Asian Crisis).
  - High savings and rapid growth enabled avoidance of inflationary debt financing despite substantial fiscal deficits in some economies.
  - Continued institutional strengthening, financial development, and in many cases more open trade policies.
- Remaining challenges:
  - Almost 700 million Asians, or 20 percent of the total population, still live in extreme poverty, with a substantial proportion in rural areas.

### III. Sources of growth — decomposition framework and principal quantitative results (1970–2005)
- Decomposition framework:
  - Output per capita = labor productivity (output per worker) × participation rates (labor force/working-age population, ages 15–64) × age structure (working-age population/total population).
  - Labor productivity decomposed into capital deepening, rising labor quality, and growing TFP.
  - Capital share in income assumed equal to 0.35; physical capital estimates based on Nehru and Dhareshwar (1993) updates; human capital from Barro and Lee (2000).
- Main quantitative findings:
  - Growth differences—both across regions and within Asia—were driven mainly by labor productivity.
  - Demographic developments provided an important boost; in some countries (Indonesia, Korea, Taiwan Province of China) the demographic impact amounted to more than 1 percentage point per year.
  - Physical capital accumulation contributed 1.75 to 3 percentage points to growth in fast-developing Asian countries.
  - Rising education levels boosted Asian growth on average by ½ percentage point.
  - TFP contributed 0.75 to 2 percentage points to growth in India, Japan, the NIEs, and Thailand.
  - Asia’s catch-up with advanced economies largely reflected capital accumulation.
  - Note: Japan’s TFP growth declined steadily after the initial takeoff (text notes “0.2 percentage points” in context).

### IV. ICT, investment, and productivity (1989–2005 dataset)
- Dataset: smaller cross-country dataset covering 1989–2005 (updated Jorgenson and Vu, 2005).
- Key empirical findings:
  - Economy-wide investment in ICT capital contributed about ½ percentage point to growth in the NIEs and China (Figure 4).
  - ICT-producing sectors account for 10 percent or more of total value added in several Asian countries, including Korea, Malaysia, the Philippines, Singapore, and Taiwan Province of China.
  - The impact of non-ICT capital accumulation is much larger than ICT’s impact.
- Caveat:
  - Estimates for TFP growth in China may be influenced by inaccurate investment price deflators.

### V. Impact of the Asian Crisis on growth and investment
- Growth rates typically recovered to pre-crisis levels.
- Investment rates in the most severely impacted countries remain below pre-crisis levels (IMF, 2005b).
- TFP post-crisis: increases may be playing a more important role, but it is still too early to detect a statistically significant post-crisis shift in trend TFP growth.
- Measurement warning: capital stock data have severe measurement problems; effective write-off of capital after mid-1990s financial crises may not be fully reflected in national accounts (relevant for ASEAN-4).

### VI. Sectoral effects — cross-sector shifts versus within-sector growth
- Definitions:
  - Sectoral reallocation effect: labor moving from low-productivity to high-productivity sectors raises aggregate labor productivity.
  - Sectoral composition effect: a higher value added share of intrinsically high-productivity-growth sectors raises aggregate productivity growth.
- Aggregation levels:
  - Broad sectors: agriculture, industry, services (World Bank WDI).
  - Manufacturing subsectors: high-skill versus low-skill (UNIDO Industrial Database).
- Focus: labor productivity rather than TFP due to limited sectoral capital stock data.

### VII. Across agriculture, industry, and services — composition and trends
- Sectoral composition:
  - Asia: relatively high share of value added in industry and a low share in services (Figure 5).
  - China and the ASEAN-4: exceptionally high industry value-added shares and low services shares; India shows the opposite pattern.
- Employment shares and productivity:
  - Developing Asia (notably China and India) have much higher employment share in agriculture and lower share in services than predicted by fundamentals.
  - Resulting pattern: relatively low agricultural productivity across developing Asia; relatively high industry productivity in China and the ASEAN-4; relatively high services productivity in India.
- Historical shifts:
  - Employment share of agriculture declined by an average 0.6 percentage point per year.
  - Movement of labor into services was at least as large as toward industry across developing Asia.

### VIII. Sectoral productivity levels and growth patterns
- Intersectoral differentials:
  - Globally, nonagricultural labor productivity ≈ three times agricultural productivity; in Asia the differential is larger.
  - Intersectoral productivity differentials remained high and widened over time in China and India.
- Productivity growth (1980–2004 pattern):
  - Globally, productivity growth in industry and agriculture exceeded that in services.
  - In Asia, productivity growth was highest in industry (except India, where services grew most rapidly).
  - After initial takeoff, productivity growth eventually decelerated, especially in services; this deceleration has not yet begun in China nor India.
  - Agriculture: little catch-up observed since the end of the Green Revolution.

### IX. Sectoral decomposition magnitudes and interpretation
- Empirical magnitudes:
  - Sectoral shifts boosted productivity growth in Asia relative to the United States by ½ percentage point per year, out of a total observed differential of 2 percentage points.
  - Regression evidence: a 1 percentage point reduction in the average annual change in the agricultural employment share is associated with a 1.5 percentage points increase in average annual labor productivity growth (after controls).
  - All Asian subregions except Japan benefited substantially over the last three decades from sectoral shifts, especially China.
  - Sectoral shift and composition effects account for about 40 percent of Asia’s productivity catch-up toward U.S. levels; about 60 percent reflects within-sector productivity growth.
  - Sectoral composition effect: positive but modest for Asia and Latin America; especially large for China and the ASEAN-4 due to very high industry shares.

### X. Within manufacturing — skill intensity, shifts, and contributions
- Manufacturing classification:
  - 28 UNIDO subsectors aggregated into 14 skill-intensive and 14 non-skill-intensive sectors; skill intensity based on income share of skilled labor.
- Cross-regional findings:
  - Asia (NIEs, China, India) has a relatively large share of manufacturing value added and employment in skill-intensive sectors compared to Latin America and predicted levels.
- Dynamics and contributions:
  - Since the mid-1960s (China reliable only since 1990), labor movement from nonskill to skill-intensive sectors ≈ the United States and much higher than Latin America.
  - Productivity levels and growth are higher in skill-intensive than nonskill-intensive sectors.
  - Most of Asia’s catch-up in manufacturing productivity attributable to high productivity growth within skill-intensive sectors; sectoral shifts contributed little.
  - Sectoral composition contribution for Asia in manufacturing was negative and close to ½ percentage point per year, driven mostly by Indonesia and Other Asia.

### XI. Policy determinants of productivity growth — variables, methods, and key regression results
- Variables analyzed:
  - Trade openness: fraction of years considered open (Welch-Wacziarg, 2003).
  - Financial development: private credit by deposit money banks and other financial institutions as percent of GDP (Beck, Demirgüç-Kunt, Levine, 2000).
  - Education: average schooling years (Barro and Lee, 2000).
  - Institutional quality: Kaufmann and Kraay government effectiveness index; Cato Institute Index of Economic Freedom (decadal regressions).
  - Cost of starting a business: share of per capita income (World Bank Doing Business).
- Methods and samples:
  - Period: 1965–2005.
  - Cross-country OLS (weighted least squares, White standard errors) with 50 countries (Tables 1–2).
  - Decade-average SUR regressions (1975–2004; 27 countries; 81 observations) with coefficients constrained equal across decades (Tables 3–4).
- Robust empirical findings (selected coefficients preserved exactly as reported):
  - Convergence (initial productivity):
    - Table 1 (1) Productivity Growth: -0.021*** (0.003)
    - Table 2 (1) Productivity Growth: -0.018*** (0.0023)
    - Table 3 (1) Productivity Growth: -0.019*** (0.0057)
    - Table 4 (1) Productivity Growth: -0.018*** (0.0058)
  - Trade openness:
    - Table 1 (1): 0.025*** (0.0045)
    - Table 2 (1): 0.020*** (0.0058)
    - Table 3 (1): 0.019*** (0.0049)
    - Table 4 (1): 0.019*** (0.0050)
    - Labor shift out of agriculture (Table 1 (5)): 0.50*** (0.12); Table 2 (5): 0.54*** (0.13); Table 3 (5): 0.26*** (0.10); Table 4 (5): 0.28*** (0.10)
  - Cost of starting a business (Table 2 coefficients):
    - (1) -0.12*** (0.028)
    - (2) -0.076** (0.035)
    - (3) -0.18** (0.087)
    - (4) -0.13** (0.054)
  - Initial financial development (selected):
    - Table 1 (1): 0.0044** (0.0019)
    - Table 1 (5): 0.093** (0.037)
    - Table 3 (1): 0.0072** (0.0029)
    - Table 4 (1): 0.0060* (0.0031)
  - Human capital:
    - Table 1 (3) Industry Labor Productivity Growth: 0.0039*** (0.0011)
    - Decade SUR Table 3 (3): 0.0059*** (0.0017)
    - Growth of human capital and labor shift out of agriculture:
      - Table 1 (5): 6.6* (3.5)
      - Table 2 (5): 7.6* (4.0)
      - Table 3 (5): 6.2** (3.1)
      - Table 4 (5): 5.7* (3.2)
  - Institutional quality / government effectiveness:
    - Table 2 (3) Government Effectiveness: 0.013*** (0.0033)
    - Table 2 (4) Government Effectiveness: 0.012*** (0.0026)
    - Cato Index proxy in decadal regressions: Table 4 (1) 0.0028 (0.0021) [not uniformly significant]
  - Agricultural employment (nonlinear effects):
    - Table 1 (1) Initial Agricultural Employment: -0.00027** (0.00011)
    - Table 1 (3) Initial Agricultural Employment: 0.030*** (0.0037)
    - Table 1 (5) Initial Agricultural Employment Squared: -0.00024*** (0.000049)
    - Decade SUR Table 3 (5) Initial Agricultural Employment Squared: -0.00046*** (0.000083)
- Interpretation highlights:
  - Trade openness and financial sector development are strongly significant determinants of aggregate and within-sector productivity growth and stimulate employment shifts out of agriculture.
  - Initial education levels are most significant for within-sector productivity growth.
  - Cost of starting a business has an economically and statistically significant negative impact on productivity growth in all regressions where included.
  - Institutional quality significantly impacts sectoral-level productivity growth; inclusion of institutions weakens significance of openness and initial schooling in some specifications.
  - Interaction evidence: financial development matters when it reduces negative effects of start-up costs.
  - Some variables (macro policy quality, FDI) are not significant once main determinants are controlled for.

### XII. The Asian story — comparative performance and policy lessons
- Comparative strengths explaining Asia’s productivity performance:
  - Asia outperforms Latin America and other developing regions on institutional quality, trade openness, and financial sector development.
  - These factors explain about 40 percent of the positive productivity growth differential between Asia and other developing economies, and reduce Asia’s advantage over advanced economies by half.
  - Trade openness largely stimulates employment shifts from agriculture; institutional quality and ease of starting a business mainly affect within-sector productivity growth.
- Regional heterogeneity:
  - Trade liberalization supported structural transformation in Japan, the NIEs, and the ASEAN-4; relatively low openness in China and India slowed the process.
  - Financial development promoted labor movement toward industry and services in Japan and the NIEs, less so elsewhere (especially India).
  - Institutional quality and ease of starting a business stimulated within-sector productivity growth in Japan and the NIEs; weaker performance in India and Other Asia dragged down productivity.
- Sectoral priorities and recommended reforms:
  - Services productivity growth has lagged and in many cases stalled convergence to advanced-economy levels.
  - Policy recommendations for services: deregulation; greater foreign competition; promote competition in infrastructure-related services (telecommunications); further opening of retail and financial sectors; lift entry restrictions in social services (health and education); increase regulatory transparency and consistency; streamline administrative procedures (including harmonizing regulations across states where decentralized, as in India).
  - Corporate governance improvements are important for growth and productivity, especially for industries dependent on external finance.

### XIII. Conclusions and policy implications — priorities to sustain convergence
- Key findings:
  - Growth in most of Asia has benefited from rapid increases in TFP and fast accumulation of physical and human capital.
  - These developments reflect a stronger institutional and policy environment (financial development, business climate, and in many cases trade openness) than in other developing economies.
  - Sectoral shifts (especially out of agriculture) and movement within manufacturing toward higher value-added contributed importantly, but about 60 percent of Asia’s productivity catch-up reflects within-sector productivity growth.
- Policy priorities:
  - Raise educational standards to support skill- and innovation-based industries and move up the value-added chain.
  - Strengthen corporate and financial-sector governance; broaden and deepen financial systems (for example, develop corporate bond markets) to finance infrastructure improvements.
  - Increase trade openness, ensure widespread access to education and health care, and encourage entrepreneurship to sustain labor reallocation away from agriculture.
  - Ensure structural flexibility (including labor markets) while establishing effective social safety nets.
  - Promote competition and productivity growth in services by removing entry barriers, streamlining regulations, and strengthening human capital.

*Italic: Source: _wp07130 - References (IMF PDF)*

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

### _wp07130 - References

### Tables and Figures Inventory
- Tables
  - 1. Determinants of Productivity Growth and Intersectoral Labor Shifts: Cross-Country OLS Regressions, Excluding Institutions and Investment Climate
  - 2. Determinants of Productivity Growth and Intersectoral Labor Shifts: Cross-Country OLS Regressions, Including Institutions and Investment Climate
  - 3. Determinants of Productivity Growth and Intersectoral Labor Shifts: Cross-Country Decade Average SUR Regressions, Excluding Institutions and Investment Climate
  - 4. Determinants of Productivity Growth and Intersectoral Labor Shifts: Cross-Country Decade Average SUR Regressions, Including Institutions and Investment Climate
- Figures
  - 1. Output Per Capita
  - 2. Selected Indicators
  - 3. Growth Decompositions
  - 4. Information and Communications Technologies (ICT) Investment and Labor Productivity Growth, 1989–2005
  - 5. Sectoral Shares of Value Added and Employment for Asia
  - 6. Sectoral Shares of Employment Over Time
  - 7. Relative Labor Productivity of Non-Agricultural Sectors Over Time
  - 8. Productivity Growth by Sector
  - 9. Sectoral Productivity Growth Since Takeoff
  - 10. Contributions to Average Labor Productivity Growth Differentials with the United States
  - 11. Skill-Intensive Manufacturing Sectors: Employment and Value-Added Shares
  - 12. Employment Share of Skill Intensive Sectors Over Time
  - 13. Productivity Growth by Sector
  - 14. Contributions to Average Manufacturing Productivity Growth Differential with the United States
  - 15. Global Determinants of Productivity Growth, 1965–2005
  - 16. Regional Determinants of Productivity Growth, 1965–2005
  - 17. Detailed Contributions to Productivity Growth

### I. INTRODUCTION — Scope and Questions
- Focus
  - Examines relative growth performance across Asia at aggregate and sectoral levels.
  - Analyzes determinants of resource shifts from low- to high-productivity sectors.
- Key questions
  - Whether development paths blazed by Japan and the NIEs are being followed by ASEAN-4, China, India, and newly emerging economies such as Vietnam.
  - Whether systematic differences exist between East Asia and the rest of Asia, or between Asia and other regions.
  - Sources of growth differences within Asia and relative to other regions, and the role of policies.
  - How Asia’s high growth rates can be sustained and whether post-Asian-crisis reforms have had detectable impacts on growth and productivity.

### II. ASIA’S ECONOMIC SUCCESS — Stylized Facts and Comparative Patterns
- Long-run performance
  - Asia’s real income per capita rose sevenfold between 1950 and 2005.
  - Significant reduction in the gap relative to the United States; contrast with Latin America and other developing economies that failed to catch up.
- Within-region heterogeneity
  - Wide disparities in per capita income: example range cited from over $33,000 in Singapore to $2,000 in Bangladesh.
  - Presence of early and late developers; later developers (including China) appear to have started takeoff at lower income levels than Japan or the NIEs.
  - Takeoff dating used in the paper: 1955 for Japan; 1967 for the NIEs; 1973 for ASEAN-4; 1979 for China; 1982 for India; 1990 for Other Asia.
- Trade, education, and demographics
  - Asia’s share of world trade more than doubled during 1970–2005; China’s rapid trade expansion stands out.
  - Rapid increase in educational attainment over the last half century.
  - Declining dependency ratios supported growth but not significantly more than in other developing regions.
  - Demographic prospects: NIEs and China likely to see rising dependency ratios within the next five years; India’s demographic transition started only relatively recently.
- Policy and institutional environment
  - Fast developers characterized by broadly stable macroeconomic environments and contained inflation (exceptions: oil shocks and the 1997 Asian Crisis).
  - High savings and rapid growth in some economies enabled avoidance of inflationary debt financing despite substantial fiscal deficits.
  - Continued institutional strengthening, financial development, and in many cases more open trade policies.
- Remaining challenges
  - Almost 700 million Asians, or 20 percent of the total population, still live in extreme poverty, with a substantial proportion in rural areas (Chen and Ravallion, 2004).

### III. PERSPIRATION OR INSPIRATION? — Sources of Growth and Decomposition Results
- Analytical decomposition framework
  - Output per capita decomposed into:
    - (1) labor productivity (output per worker);
    - (2) participation rates (ratio of labor force to working-age population, where working age = ages 15–64 inclusive);
    - (3) age structure of the population (ratio of working-age population to total population, or one minus the total dependency ratio).
  - Labor productivity further decomposed into:
    - (1) capital deepening (increases in physical capital per worker);
    - (2) rising labor quality;
    - (3) growing TFP.
  - Methodological bases referenced: Jorgenson, Ho, and Stiroh, 2005; Jorgenson, forthcoming; physical capital estimates based on Nehru and Dhareshwar, 1993, updated as in Fajnzylber and Lederman, 1999, using World Economic Outlook data on gross fixed capital formation; human capital based on Barro and Lee, 2000. Capital share in income assumed equal to 0.35.
- Main quantitative findings (1970–2005)
  - Growth differences—both across regions and within Asia—were driven mainly by labor productivity.
  - Demographic developments provided an important boost to growth in both Asia and Latin America; in some countries (Indonesia, Korea, Taiwan Province of China), the demographic impact amounted to more than 1 percentage point per year.
  - Asia enjoyed both faster physical capital accumulation and faster TFP growth than other developing economies.
  - Asia’s catch-up with advanced economies largely reflected capital accumulation.
  - Physical capital accumulation contributed 1.75 to 3 percentage points to growth in fast-developing Asian countries.
  - Rising education levels boosted Asian growth on average by ½ percentage point.
  - TFP contributed 0.75 to 2 percentage points to growth in India, Japan, the NIEs, and Thailand.
  - Results for the NIEs broadly similar to Young (1995) over comparable periods, except TFP growth for Singapore through 1990 is estimated at over 1 percentage point (text truncated here).

### Policy Implications Highlighted in the Text
- Factors associated with Asia’s performance
  - Greater trade openness.
  - Macroeconomic stability.
  - Financial development.
  - Educational attainment in many cases.
  - More favorable institutional and policy environment than observed in other developing economies.
- Forward-looking recommendations
  - Further improvements in policies and institutional quality would help to:
    - Sustain high sectoral productivity growth rates.
    - Facilitate continued shift of resources from agriculture to industry and services.
    - Support sustained rapid growth, convergence toward advanced-economy income levels, and elimination of poverty across the region.

*Italic: Source: _wp07130 - References (IMF PDF)_*

### 0.2 percentage points. In Japan, TFP growth declined steadily after the initial takeoff. In the

### _wp07130 - 0.2 percentage points. In Japan, TFP growth declined steadily after the initial takeoff. In the

### ICT, investment, and productivity
- Dataset and period: smaller cross-country dataset covering 1989–2005 (updated version of Jorgenson and Vu, 2005).
- Key empirical findings:
  - Economy-wide investment in ICT capital contributed about ½ percentage point to growth in the NIEs and China (Figure 4).
  - ICT-producing sectors account for 10 percent or more of total value added in several Asian countries, including Korea, Malaysia, the Philippines, Singapore, and Taiwan Province of China; however, it was not possible to estimate TFP growth within these sectors.
  - Asia does not stand out on the ICT-impact dimension; the impact of non-ICT capital accumulation is much larger.
- Caveats: estimates for TFP growth in China may be influenced by inaccurate investment price deflators.

### Impact of the Asian Crisis on growth and investment
- Growth rates: growth rates have typically recovered to pre-crisis levels.
- Investment rates: in the countries most severely impacted, investment rates are still below pre-crisis levels (IMF, 2005b).
- TFP post-crisis: increases in TFP may be playing a more important role, but it is still too early to detect any statistically significant post-crisis shift in trend TFP growth.
- Measurement warning: available data are plagued by severe measurement problems, especially with respect to capital stocks; e.g., effective write-off of capital after mid-1990s financial crises may not be fully reflected in national accounts, relevant for the ASEAN-4.

### Sectoral effects: cross-sector shifts versus within-sector growth
- Definitions:
  - Sectoral reallocation effect: labor moving from low-productivity to high-productivity sectors raises aggregate labor productivity.
  - Sectoral composition effect: a higher value added share of sectors with intrinsically high productivity growth raises aggregate productivity growth.
- Aggregation levels analyzed:
  - Broad sectors: agriculture, industry, services (World Bank, 2006 WDI).
  - Manufacturing subsectors: high-skill versus low-skill (UNIDO Industrial Database).
- Focus: labor productivity rather than TFP due to limited data on sectoral capital stocks.

### Across agriculture, industry, and services — composition and trends
- Asia’s sectoral composition:
  - Asia has a relatively high share of value added in industry and a low share in services (Figure 5).
  - Japan and the NIEs resemble similarly placed advanced economies; China and the ASEAN-4 have an exceptionally high share of value added in industry and low share in services; India shows the opposite pattern.
- Employment shares and productivity:
  - Developing Asia, notably China and India, have a much higher employment share in agriculture and correspondingly lower share in services than predicted by fundamentals.
  - Result: relatively low agricultural productivity throughout developing Asia; productivity relatively high in industry for China and the ASEAN-4 and in services for India.
- Historical shifts:
  - The employment share of agriculture declined by an average 0.6 percentage point per year.
  - Example: agriculture accounted for about a third of Korea’s and Taiwan Province of China’s economies in the 1960s, but less than one-tenth by the 1980s.
  - Movement of labor into services was at least as large as toward industry across developing Asia.

### Sectoral productivity levels and growth patterns
- Intersectoral differentials:
  - For the world, labor productivity in nonagricultural sectors is about three times higher than in agriculture; in Asia the differential is even larger (Figure 7).
  - Intersectoral productivity differentials remained high and widened over time in China and India.
- Productivity growth by sector (1980–2004 pattern):
  - Productivity growth in both industry and agriculture exceeded that in services globally.
  - In Asia:
    - Productivity growth was highest in industry (except India, where services grew most rapidly).
    - Productivity growth in industry and (until recently) services far exceeded that in other regions, implying catch-up toward U.S. levels.
    - After the initial takeoff, productivity growth eventually decelerated, especially in services—this deceleration has not yet begun in China nor India (Figure 9).
  - Agriculture: little catch-up observed since the end of the Green Revolution.

### Sectoral decomposition of productivity growth — methodology and magnitudes
- Decomposition components:
  - Sectoral reallocation effect (employment-share changes weighted by sector productivity).
  - Within-sector productivity growth (sector productivity growth weighted by initial value added share).
  - Cross-country addition: sectoral composition effect (difference in sector value added shares relative to comparator, e.g., the United States).
- Empirical results:
  - Sectoral shifts boosted productivity growth in Asia relative to the United States by ½ percentage point per year, out of a total observed differential of 2 percentage points.
  - Regression evidence: a 1 percentage point reduction in the average annual change in the agricultural employment share is associated with a 1.5 percentage points increase in average annual labor productivity growth (after controls).
  - All Asian subregions except Japan benefited substantially over the last three decades from sectoral shifts, especially China.
  - Altogether, sectoral shift and composition effects account for about 40 percent of Asia’s productivity catch-up toward U.S. levels; the greater part reflects strong within-sector productivity growth.
  - Sectoral composition effect: positive but modest for Asia and Latin America; especially large for China and the ASEAN-4 due to very high industry shares.

### Within manufacturing: skill intensity, shifts, and contributions
- Manufacturing subsector classification: 28 UNIDO subsectors aggregated into 14 skill-intensive and 14 non-skill-intensive sectors; skill intensity based on income share of skilled labor (input-output matrix for South Africa).
- Cross-regional comparisons:
  - Asia, particularly the NIEs, China, and India, have a relatively large share of manufacturing value added and employment in skill-intensive sectors compared to Latin America and predicted levels.
- Dynamics:
  - Since the mid-1960s (reliable China data only since 1990), the rate of labor movement from nonskill to skill-intensive sectors has been about the same as in the United States and much higher than in Latin America.
  - Both productivity levels and productivity growth are higher in skill-intensive than in nonskill-intensive sectors.
- Contributions to manufacturing productivity:
  - Aggregate manufacturing productivity grew faster in Asia than the United States, but the differential was smaller than for overall productivity; ASEAN-4 and India had manufacturing productivity growth slower than the United States.
  - Most of Asia’s catch-up in manufacturing productivity was attributable to high productivity growth within skill-intensive sectors; sectoral shifts contributed little.
  - Sectoral composition contribution for Asia in manufacturing was actually negative and close to ½ percentage point per year, driven mostly by Indonesia and Other Asia.

### Policy determinants of productivity growth — variables examined and main results
- Policy and institutional variables analyzed:
  - Trade policy stance: fraction of years considered open (Welch-Wacziarg, 2003).
  - Financial sector development: private credit by deposit money banks and other financial institutions as percent of GDP (Beck, Kunt, Levine, 2000).
  - Education levels: average schooling years (Barro and Lee, 2000).
  - Institutional quality: Kaufmann and Kraay (2005) government effectiveness index (cross-sectional regressions) and Cato Institute Index of Economic Freedom (decadal regressions).
  - Cost of starting a business: share of per capita income (World Bank Doing Business).
- Estimation samples and methods:
  - Dataset period: 1965–2005.
  - Approaches: cross-country regressions using weighted least squares (robust standard errors); decadal SUR regressions for subsets.
- Robust empirical findings:
  - Convergence effect: negative and significant coefficient on initial productivity in all regressions.
  - Initial share of agricultural employment predicts slower growth and is a major determinant of magnitude of labor shifts.
  - Trade openness and financial sector development are strongly significant determinants of aggregate and within-sector productivity growth and stimulate employment shifts out of agriculture.
  - Initial education levels are most significant for within-sector productivity growth.
  - Cost of starting a business has an economically and statistically significant negative impact on productivity growth in all regressions; inclusion reduces significance of financial development.
  - Institutional quality significantly impacts productivity growth at the sectoral level; controlling for institutions weakens significance of openness and initial schooling.
  - Interaction: financial development matters to the extent that it reduces the negative effects of start-up costs (in specifications allowing interaction terms).
- Robustness and caveats:
  - Time-varying institutional measures (Cato Index) performed poorly in decadal regressions.
  - No significant effect of institutional quality on intersectoral labor shifts (coefficient had wrong sign).
  - Some variables (macro policy quality, FDI) not significant once main determinants controlled for.

### The Asian story — comparative performance and policy lessons
- Asia’s comparative strengths:
  - Asia outperforms Latin America and other developing regions on institutional quality, trade openness, and financial sector development (Figure 16).
  - These factors explain about 40 percent of the positive productivity growth differential between Asia and other developing economies, and reduce Asia’s advantage over advanced economies by half (Figure 17).
  - Trade openness largely stimulates employment shifts from agriculture; institutional quality and ease of starting a business mainly affect within-sector productivity growth.
- Regional heterogeneity:
  - Trade liberalization supported structural transformation in Japan, the NIEs, and the ASEAN-4; relatively low openness in China and India slowed the process.
  - Financial development promoted labor movement toward industry and services in Japan and the NIEs, less so elsewhere (especially India).
  - Institutional quality and ease of starting a business stimulated within-sector productivity growth in Japan and the NIEs; weaker performance in India and Other Asia dragged down productivity.
- Sectoral priorities and reforms:
  - Services productivity growth has lagged and, in many cases, stalled convergence to advanced-economy levels.
  - Policy recommendations for services sectors include deregulation, greater foreign competition, promoting competition in infrastructure-related services (telecommunications), further opening retail and financial sectors, lifting entry restrictions in social services (health and education), increasing regulatory transparency and consistency, and streamlining administrative procedures (including harmonizing regulations across states where regulation is decentralized, as in India).
  - Corporate governance improvements are important for growth and productivity, especially for industries dependent on external finance; Asia still lags advanced-economy standards.
- Country-specific notes:
  - China’s productivity growth strongly linked to post-1979 reforms: property rights development (notably in agriculture), market opening, removal of barriers to capital and labor mobility, and Special Economic Zones.
  - Slow TFP growth in the ASEAN-4, especially the Philippines, may reflect weaknesses in institutional quality and infrastructure.

### Conclusions and policy implications
- Key findings:
  - Growth in most of Asia has benefited from rapid increases in TFP and fast accumulation of physical and human capital.
  - These developments reflect a stronger institutional and policy environment (financial development, business climate, and in many cases trade openness) than in other developing economies.
  - Sectoral shifts (especially out of agriculture) and movement within manufacturing toward higher value-added contributed importantly, but about 60 percent of Asia’s productivity catch-up reflects within-sector productivity growth.
- Policy priorities to sustain convergence:
  - Foster higher standards of education to support skill- and innovation-based industries and move up the value-added chain.
  - Continue strengthening corporate and financial-sector governance; broaden and deepen financial systems (e.g., develop corporate bond markets) to finance infrastructure improvements.
  - Increase trade openness, ensure widespread access to education and health care, encourage entrepreneurship to sustain labor reallocation away from agriculture.
  - Ensure structural flexibility (including labor markets) while establishing effective social safety nets.
  - Promote competition and productivity growth in services by removing barriers to entry, streamlining regulations, and strengthening human capital.

*Source: IMF working paper content (pages excerpted).*

### REFERENCES

### _wp07130 - REFERENCES

### Economic growth, productivity, and development
- Barro, Robert J., 1997, “Determinants of Economic Growth: A Cross-Country Empirical Study,” (Cambridge, Massachusetts: MIT Press).
- Barro, Robert J., and Jong-Wha Lee, 2000, “International Data on Educational Attainment: Updates and Implications,” CID Working Paper No. 42 (Cambridge, Massachusetts: Center for International Development at Harvard University).
- Barro, Robert J., 2001, “Economic Growth in East Asia Before and After the Financial Crisis,” NBER Working Paper No. 8330 (Cambridge, Massachusetts: National Bureau of Economic Research).
- Bloom, David E., David Canning, and Pia N. Malaney, 1999, “Demographic Change and Economic Growth in Asia,” CID Working Paper No. 15 (Cambridge, Massachusetts: Center for International Development at Harvard University).
- Bosworth, Barry, and Susan M. Collins, 2003, “The Empirics of Growth: an Update,” Brookings Papers on Economic Activity: 2, Brookings Institution.
- Denison, Edward F., 1962, “The Sources of Economic Growth in the United States and the Alternatives Before Us,” Supplementary Paper of the Committee for Economic Development No. 13 (New York).
- Denison, Edward F., 1967, “Why Growth Rates Differ; Postwar Experience in Nine Western Countries,” (Washington: Brookings Institution).
- Dekle, Robert, and Guillaume Vandenbroucke, 2006, “A Quantitative Analysis of China’s Structural Transformation,” (unpublished; University of Southern California). Available via the Internet at: http://ssrn.com/abstract=902497.
- Edwards, Sebastian, 1998, “Openness, Productivity, and Growth: What Do We Really Know?” Economic Journal, Vol. 108 (March), pp. 383–98.
- Fajnzylber, Pablo, and Daniel Lederman, 1999, “Economic Reforms and Total Factor Productivity Growth in Latin America and the Caribbean (1950–95). An Empirical Note,” Policy Research Working Paper No. 2114 (Washington: World Bank).
- Fukao, Kyoji, Tomohiko Inui, Hiroki Kawai, and Tsutomu Miyagawa, 2003, “Sectoral Productivity and Economic Growth in Japan: 1970–98: An Empirical Analysis Based On the JIP Database,” ESRI Cabinet Office, Discussion Paper Series No. 67 (Tokyo: Economic and Social Research Institute).
- Gupta, Poonam, 2005, “Understanding the Growth Momentum in India’s Services,”

### Corporate governance, finance, and accounting
- Becht, Marco, Patrick Bolton, and Ailsa Roell, 2003, “Corporate Governance and Control,” in Handbook of the Economics of Finance, ed. by George M. Constantinides, Milton Harris, and Rene M. Stulz, Vol. 1, pp. 1–109 (Amsterdam, Boston: Elsevier/North-Holland).
- Beck, Thorsten, Asli Demigüç-Kunt, and Ross Levine, 2000, “A New Database on Financial Development and Structure,” World Bank Economic Review 14, pp. 597–605 (Washington: World Bank).
- Berglöf, Erik, and Stijn Claessens, 2006, “Enforcement and Good Corporate Governance in Developing Countries and Transition Economies,” The World Bank Research Observer, Vol. 21, No.1 (Spring), pp. 121–50.
- Bhattachrya, Utpal, Hazem Daouk, and Michael Welker, 2003, “The World Price of Earnings Opacity,” Accounting Review, Vol. 78 (July), pp. 641–78.
- Cheung, Stephen, and Hasung Jang, forthcoming, “PECC Macro Corporate Governance Scorecard Project: Evaluation of Corporate Governance in East Asian Economies.”
- Coles, Jeffrey, Michael Lemmon, and Felix Meschke, forthcoming, “Structural Models and Endogeneity in Corporate Finance: The Link Between Managerial Ownership and Performance,” Journal of Financial Economics.
- De Nicolò, Gianni, Luc Laeven, and Kenichi Ueda, 2006, “Corporate Governance Quality in Asia: Comparative Trends and Impact,” (unpublished; Washington: International Monetary Fund, Research Department).
- CIFAR, 1993, International Accounting and Auditing Trends, 3rd edition (Princeton: Center for International Financial Analysis and Research, Inc.).

### Infrastructure, regulation, and external spillovers
- Calderon, César, and Luis Serven, 2004, “The Effects of Infrastructure Development on Growth and Income Distribution,” World Bank Policy Research Working Paper No. 3400 (Washington: World Bank).
- Coe, David T., Elhanan Helpman and Alexander W. Hoffmaister, 1997, “North-South R&D Spillovers,” Economic Journal, Vol. 107 (January), pp. 134–49.
- Conway, Paul, Donato De Rosa, Giuseppe Nicoletti, and Faye Steiner, forthcoming, “Regulation, Competition and Productivity Convergence,” OECD Economics Department Working Paper (Paris: Organization for Economic Cooperation and Development).

### Regional studies, crises, and poverty
- Cerra, Valerie, and Sweta C. Saxena, 2003, “Did Output Recover from the Asian Crisis,” IMF Working Paper 03/48 (Washington: International Monetary Fund).
- Chen, Shaohua, and Martin Ravallion, 2004, “How Have theWorld's Poorest Fared Since the Early 1980s?” World Bank Research Observer, Vol. 9 No.2 (Fall), pp. 141–69.
- Dekle, Robert, and Guillaume Vandenbroucke, 2006, “A Quantitative Analysis of China’s Structural Transformation,” (unpublished; University of Southern California). Available via the Internet at: http://ssrn.com/abstract=902497.

*References as listed in _wp07130 - REFERENCES*

### Chapter II in India: Selected Issues, IMF Country Report No. 05/87 (Washington:

### Chapter II — India: Selected Issues, IMF Country Report No. 05/87 (Chapter II content unit)

### Executive summary
- Cross-country empirical analysis (1965 to 2004; repeated-decade SUR 1975–2004 for some specifications) identifies robust correlates of productivity growth and intersectoral labor shifts across countries and decades.
- Key positive correlates of productivity and labor reallocation include: trade openness; initial financial development (in some specifications); growth in human capital for some outcomes; and lower cost of starting a business (i.e., easier entry).
- Key negative correlates include: higher initial productivity (convergence effect) and higher initial share of agricultural employment (nonlinear effects via squared term).
- Government effectiveness shows positive and statistically significant associations with services labor productivity in specifications that include institutional measures.
- Sectoral evidence highlights rapid structural change in Asia since country-specific growth takeoff years (Japan 1955; NIEs 1967; ASEAN-4 1973; China 1979; India 1982; other Asia 1990) with ICT investment and capital deepening important for labor productivity growth.

### Cross-country OLS regressions — Main coefficient findings (Tables 1 and 2)
- Dependent variables covered: (1) Productivity Growth (aggregate); (2) TFP Growth; (3) Industry Labor Productivity Growth; (4) Services Labor Productivity Growth; (5) Labor Shift Out of Agriculture.
- Table 1 (excluding institutions and investment climate) — selected coefficients (coefficient with White standard error in parentheses; significance markers retained):
  - Initial Productivity:
    - (1) Productivity Growth: -0.021*** (0.003)
    - (2) TFP Growth: -0.0081*** (0.0025)
    - (3) Industry Labor Productivity Growth: -0.027*** (0.0043)
    - (4) Services Labor Productivity Growth: -0.021*** (0.0049)
  - Initial Human Capital:
    - (1) 0.0015** (0.00065)
    - (3) 0.0039*** (0.0011)
    - (4) 0.0027*** (0.0010)
  - Trade Openness:
    - (1) 0.025*** (0.0045)
    - (2) 0.023*** (0.006)
    - (3) 0.027*** (0.0089)
    - (4) 0.019* (0.010)
    - (5) 0.50*** (0.12)
  - Initial Financial Development:
    - (1) 0.0044** (0.0019)
    - (3) 0.0073** (0.0031)
    - (4) 0.0055** (0.0027)
    - (5) 0.093** (0.037)
  - Initial Agricultural Employment:
    - (1) -0.00027** (0.00011)
    - (2) -0.00018*** (0.000081)
    - (3) 0.030*** (0.0037)
    - Initial Agricultural Employment Squared (affecting labor shift out of agriculture):
    - (5) -0.00024*** (0.000049)
  - Growth of Human Capital:
    - (5) 6.6* (3.5)
  - Sample size and fit:
    - Number of countries: 50 across columns (1)–(5)
    - R-squared: 0.71 (1), 0.47 (2), 0.44 (3), 0.34 (4), 0.81 (5)

- Table 2 (including institutions and investment climate) — selected coefficients:
  - Initial Productivity:
    - (1) -0.018*** (0.0023)
    - (2) -0.0060** (0.0024)
    - (3) -0.023*** (0.0042)
    - (4) -0.021*** (0.0034)
  - Initial Human Capital:
    - (2) -0.0012* (0.00068)
  - Trade Openness:
    - (1) 0.020*** (0.0058)
    - (2) 0.019* (0.0084)
    - (3) 0.010 (0.010)
    - (4) 0.0044 (0.0078)
    - (5) 0.54*** (0.13)
  - Initial Financial Development:
    - (1) 0.0013 (0.0015)
    - (5) 0.088** (0.036)
  - Cost of Starting a Business (percent of income per capita; World Bank Doing Business database):
    - (1) -0.12*** (0.028)
    - (2) -0.076** (0.035)
    - (3) -0.18** (0.087)
    - (4) -0.13** (0.054)
    - (5) -1.1 (1.4) [not significant]
  - Government Effectiveness (Kaufmann, Kraay, and Mastruzzi 2005):
    - (3) 0.013*** (0.0033)
    - (4) 0.012*** (0.0026)
  - Initial Agricultural Employment:
    - (1) -0.00017** (0.000081)
    - (3) 0.029*** (0.0044)
    - Initial Agricultural Employment Squared:
    - (5) -0.00022*** (0.000061)
  - Growth of Human Capital:
    - (5) 7.6* (4.0)
  - Sample size and fit:
    - Number of countries: 50
    - R-squared: 0.80 (1), 0.54 (2), 0.68 (3), 0.66 (4), 0.82 (5)

### Decade-average SUR regressions — main findings (Tables 3 and 4)
- Tables 3 and 4 repeat the analysis using repeated cross-sectional decade-average SUR regressions (decades constrained equal across periods; 1975–2004; 27 countries; 81 observations).
- Table 3 (excluding institutions) — selected coefficients:
  - Initial Productivity:
    - (1) -0.019*** (0.0057)
    - (2) -0.012** (0.0048)
    - (3) -0.032*** (0.0063)
    - (4) -0.026*** (0.0056)
  - Initial Human Capital:
    - (3) 0.0059*** (0.0017)
    - (4) 0.0024* (0.0014)
  - Trade Openness:
    - (1) 0.019*** (0.0049)
    - (2) 0.019*** (0.0044)
    - (3) 0.020*** (0.0078)
    - (4) 0.015** (0.0065)
    - (5) 0.26*** (0.10)
  - Initial Financial Development:
    - (1) 0.0072** (0.0029)
    - (2) 0.0065** (0.0027)
    - (4) 0.011*** (0.0038)
  - Initial Agricultural Employment Squared:
    - (5) -0.00046*** (0.000083)
  - Growth of Physical Capital / Labor:
    - (5) 1.7* (0.93)
  - Growth of Human Capital:
    - (5) 6.2** (3.1)
  - Sample: Number of countries 27; observations 81; R-squared 0.23 (1), 0.21 (2), 0.37 (3), 0.36 (4), 0.63 (5)

- Table 4 (including institutions) — selected coefficients:
  - Initial Productivity:
    - (1) -0.018*** (0.0058)
    - (2) -0.012** (0.0050)
    - (3) -0.032*** (0.0063)
    - (4) -0.026*** (0.0056)
  - Initial Human Capital:
    - (3) 0.0057*** (0.0017)
  - Trade Openness:
    - (1) 0.019*** (0.0050)
    - (2) 0.019*** (0.0044)
    - (3) 0.020*** (0.0078)
    - (4) 0.015** (0.0065)
    - (5) 0.28*** (0.10)
  - Initial Financial Development:
    - (1) 0.0060* (0.0031)
    - (2) 0.006** (0.0028)
    - (4) 0.010*** (0.0040)
  - Institutional Quality (Cato Institute Index of Economic Freedom proxy):
    - (1) 0.0028 (0.0021) [insignificant in many columns]
  - Initial Agricultural Employment Squared:
    - (5) -0.00044*** (0.000086)
  - Growth of Physical Capital / Labor:
    - (5) 1.84** (0.93)
  - Growth of Human Capital:
    - (5) 5.7* (3.2)
  - Sample: Number of countries 27; observations 81; R-squared 0.24 (1), 0.20 (2), 0.30 (3), 0.46 (4), 0.53 (5)

### Sectoral and regional patterns (figures and captions)
- Growth takeoff years used for comparative series:
  - Japan: 1955
  - Newly industrialized economies (NIEs): 1967
  - ASEAN-4 (Indonesia, Malaysia, the Philippines, and Thailand): 1973
  - China: 1979
  - India: 1982
  - Other Asian economies: 1990
- Figure-based highlights (captions and series):
  - Output per capita and share of world trade increase markedly in Asia following growth takeoff dates (sources: Maddison 2003; WEO).
  - Selected indicators (share of world trade, share of population with some schooling, dependency ratios, financial development as percent of GDP, inflation) are shown for Asia relative to other regions (sources: WDI 2006; Barro and Lee 2000; CEIC; UN population prospects).
  - ICT investment and labor productivity: Figure 4 shows growth impact of ICT capital accumulation versus non-ICT capital accumulation, 1989–2005; ICT contributes positively to labor productivity growth across regions and within Asia.
  - Sectoral shares: Figures 5–7 report shares of value added and employment by sector (Agriculture, Industry, Services) across Asia, with actual and predicted values; major structural shifts away from agriculture toward industry and services are evident in Asia since takeoff.
  - Relative labor productivity of non-agricultural sectors increases over time relative to agriculture in Asia (Figure 7).
  - Productivity growth by sector: Figures 8 and 13 show annual percent changes across Agriculture, Industry, Services, and for skill-intensive vs non-skill-intensive manufacturing sectors; Asia displays strong industry and services productivity growth compared with other regions.
  - Skill-intensive manufacturing: Figures 11 and 12 show employment and value-added shares for skill-intensive manufacturing sectors (UNIDO) and their evolution over time; Asia and many Asian subgroups show higher shares relative to other regions.
  - Regional determinants of productivity growth: Figures 15–17 present levels of determinants (trade openness, institutional quality, cost of starting a business, financial sector development, initial schooling, initial non-agricultural employment) expressed in multiples of sample standard deviations and differences from Asia averages, used to decompose productivity growth differences.

### Empirical interpretation and policy-relevant associations (based on reported results)
- Convergence effect:
  - Initial productivity has a consistently negative and statistically significant coefficient across specifications (e.g., -0.021*** in Table 1 column (1); -0.018*** in Table 2 column (1)), indicating conditional convergence in productivity across countries in the sample period.
- Openness and trade:
  - Trade openness exhibits a positive and generally strongly significant association with productivity growth and labor reallocation (e.g., 0.025*** in Table 1 column (1); 0.020*** in Table 2 column (1); decade estimates 0.019*** in Tables 3–4), highlighting the empirical role of openness in raising productivity and promoting sectoral shifts.
- Entry costs and investment climate:
  - Cost of starting a business is negatively and significantly associated with aggregate productivity growth and sectoral productivity outcomes where included (Table 2: -0.12*** column (1); -0.076** column (2); -0.18** column (3); -0.13** column (4)), indicating easier firm entry (lower cost) correlates with higher productivity growth.
- Financial development:
  - Initial financial development shows positive and sometimes significant associations with aggregate productivity and sector outcomes (e.g., Table 1 (1) 0.0044**; decade SUR Table 3 (1) 0.0072**), and is positively associated with labor shifts out of agriculture in some specifications (Table 1 (5) 0.093**).
- Human capital:
  - Initial human capital (average schooling years) yields mixed results: positive and significant for certain sectoral productivity outcomes (e.g., Table 1 (3) 0.0039***; Table 3 (3) 0.0059***), but insignificant or marginal in some specifications; growth in human capital shows positive association with labor shift out of agriculture (Table 1 (5) 6.6*; Table 2 (5) 7.6*; Table 3 (5) 6.2**; Table 4 (5) 5.7*).
- Institutional quality and government effectiveness:
  - Government effectiveness is positively associated with services productivity (Table 2 (3) 0.013***; Table 2 (4) 0.012***), while broader institutional quality measures (Cato index proxy) are not uniformly significant in decade regressions (Table 4 (1) 0.0028 (0.0021), not significant).
- Agricultural employment and structural change:
  - Initial agricultural employment and its squared term show nonlinear relationships: negative linear coefficients for aggregate productivity but positive coefficients for industry productivity and a negative squared term for labor shift out of agriculture (Tables 1–4), reflecting complex reallocation dynamics as economies transition.

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07130.pdf_