## _wp13108

## Source details

**Canonical URL:** [_wp13108](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13108.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13108.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13108.pdf.json)

---

### I. Introduction: scope and contributions
- Dataset coverage and scope:
  - Covers 178 countries over 1962–2010.
  - Provides breakdowns at the SITC 4-digit and BEC 3-digit levels.
  - More than 20 million quality estimates produced; trade dataset contains 45.3 million observations at SITC 4-digit (Revision 1) level.
  - Estimation conducted for 851 products (SITC 4-digit-plus).
- Main contributions:
  - New, extensive estimates of export quality suitable for broad cross-country and long-time analysis.
  - Stylized facts on export quality evolution, sectoral patterns, and the relationship between quality and income.
  - Initial econometric analysis of determinants of quality upgrading.

### II. Estimating product quality: methodology and data
- Core identification and estimation strategy:
  - Starts from unit values p_mxt and relates ln p_mxt to unobservable quality θ_mxt, exporter income per capita y_xt, and distance Dist_mx (equation (1) in text).
  - Specifies a quality-augmented gravity equation separately for each product (equation (2)).
  - Substitutes observable components and estimates via two-stage least squares to address endogeneity of lagged price*importer-income interaction (equation (3)).
  - Instrument: lagged value of p_mxt * ln y_it (up to 5 years back); where unit value in preceding year unavailable, use nearest preceding year up to 5 years; if none, observation excluded.
- Key dataset construction and cleaning rules:
  - Starts with COMTRADE; supplements importer-reported data with exporter-reported data where importer reports missing; U.S. export flows take precedence over importer-reported flows.
  - Uses SITC 4-digit-plus products when quantities reported in multiple unit sets; total products = 851.
  - Exclusion rules for smallness or outliers:
    - Drop SITC 4-digit-plus if: (i) comprises less than 1 percent of total observations or trade value of corresponding SITC 4-digit product; or (ii) has less than 1000 observations and comprises less than 25 percent of total observations or trade value of corresponding SITC 4-digit product.
    - Exclude observations with: (i) quantity of 1; or (ii) total trade value of less than $7,500 at 1989 prices; or (iii) a unit value above the 95th or below the 5th percentile in 1989 prices within a given product.
  - Gravity controls from CEPII include: common border, common language, preferential trade agreement indicator, colonial relationship, and common colonizer; distance is great-circle distance between importer and exporter.
  - Income per capita data from Penn World Tables, version 7.1; preferential trade agreements from WTO RTA database.
- Quality calculation and aggregation:
  - Quality θ_mxt computed by adjusting unit values for production-cost differences and selection bias from distance (equation (4)).
  - Quality and importers’ taste for quality δ not separately identified.
  - Quality estimates normalized by their 90th percentile within each product-year; resulting values typically range between 0 and 1.2.
  - Aggregations performed using current trade values as weights to SITC 4-, 3-, 2-, and 1-digit levels and to country totals; normalization to the 90th percentile repeated at each aggregation step.
  - Also produces aggregations based on BEC and three broad sectors (agriculture, non-agricultural commodities, and manufactures).

### III. Stylized facts on export quality
- Quality versus unit values:
  - Unit values are much more dispersed than quality (Figure 1).
  - Quality and unit values are correlated at lower quality levels, but once quality reaches about 85 percent of the world frontier value, quality and unit values are no longer correlated.
  - Quality evolves gradually: changes of more than 20 percent are rare in any subperiod (1962–80; 1980–95; 1995–2010).
  - Quality increases often not accompanied by unit value increases; some countries register considerable quality increases with stable unit values (e.g., computer and electronics sectors).
- Sectoral examples:
  - Passenger motor cars (SITC 7321):
    - U.S. export quality slightly below world frontier on average; fluctuations led to market share decline when quality fell in the 1980s.
    - Germany: high quality and high prices throughout.
    - Japan: strong quality upgrading through sample period; surpassed German quality in 2000s with moderate price rises → increased market share.
    - Korea: low quality until mid–1980s; significant upgrading in early 1990s and 2000s; prices remained relatively low → market share increased.
  - Apparel (SITC 84):
    - China: quality increased from 70 to 90 percent of world frontier since 1980; market share rose drastically; prices remained low at 40 percent of world frontier.
    - Bangladesh and India: strong market share increases; weaker quality increases → no price increases.
    - Italy: maintained world frontier quality but market share declined as prices rose.
    - Korea and Thailand: past increases in market share with rising quality and stable prices; later diversified away from textiles and retained higher-quality segments with falling market shares.
- Quality ladders and country positions:
  - Lengths of sectoral quality ladders vary considerably across sectors and countries (Figure 5).
  - Examples:
    - Tanzania: concentrated in primary and agricultural exports, near bottom of ladder → large potential for quality upgrading; horizontal diversification to manufactures may help.
    - Vietnam: considerable upgrading in miscellaneous manufactures but still room for further upgrading.
    - Malaysia and China: near world frontier in electronics → limited scope for further upgrading within existing exports; may require horizontal diversification or moving to more sophisticated tasks.
- Quality along the development path:
  - Income per capita correlated with export quality across aggregate, manufacturing, agriculture, and non-agricultural commodities (Figure 6).
  - Speed and thresholds:
    - Quality upgrading particularly rapid until GDP per capita reaches $10,000.
    - Convergence continues at a diminishing rate and is largely complete by GDP per capita of $20,000.
    - Among high-income countries, average export quality levels vary within a narrow band.
  - Agriculture and commodities:
    - Quality can and does increase substantially in agricultural products and commodities; this may reflect shifts to more processed products.
    - Quality ladders vary across subsectors in agriculture and manufacturing (Figure 7).
  - Robustness & subsamples:
    - Within-country changes: quality increases with income even controlling for country fixed effects.
    - Small states (population <1.5 million in 2010 or 2011): quality rises with income particularly sharply for income levels below $10,000.
    - Commodity exporters (commodities >50 percent of exports): potential for quality upgrading exists but may be constrained by exogenous factors like mineral grade.
  - Unit values increase with income at a relatively constant rate, especially for manufacturing.
- Regional and income-group patterns:
  - MICs: steady convergence toward world quality frontier in manufacturing since the 1970s; agriculture convergence started later, mid–1990s.
  - LICs: manufacturing quality improvement started in mid–1990s; agricultural quality not yet taken off; non-agricultural commodity quality has deteriorated since the 1980s.
  - East Asia: particularly fast quality upgrading, especially in manufactures; commodity quality increased in 1970s–1980s due to elementary processing; agriculture improved later (late 1980s).
  - Sub-Saharan Africa and South Asia: lagging but recent signs of convergence; manufacturing quality up sharply in sub-Saharan Africa since late 1990s and in South Asia since early 2000s; agriculture in South Asia may be taking off.
  - MENA: manufacturing quality increased 1960s–1980s then stagnated; agriculture shows no sustained increases.
  - Latin America: export quality stagnated for several decades.
  - Significant cross-country heterogeneity within regions (Figure 11); examples of fast convergers: Japan, Korea, China, Vietnam; slower or stalled convergence: Malaysia, Thailand, India, Indonesia, Bangladesh; in Africa: Uganda, Nigeria, Tanzania, South Africa improved rapidly since mid-1990s; Ghana, Kenya, Cameroon stagnant.

### IV. Determinants of quality upgrading: panel regression findings
- Empirical specification:
  - Growth rate of product quality regressed on initial product quality, initial income per capita, initial institutional quality, initial human capital, and other controls (equation (5)).
  - Variables averaged over 10-year non-overlapping periods.
  - Institutional quality measured by “constraints on the executive variable” from Polity IV (robust to Kaufmann-Kraay-Mastruzzi indicators).
  - Human capital measured by secondary-school completion rate (World Development Indicators).
- Key econometric findings:
  - Convergence:
    - Growth rate of product quality depends negatively on its initial level — unconditional and conditional convergence.
    - Speed of convergence for individual products: about 5 percent per annum (Table 3).
    - Within manufacturing: speed of convergence about 7 percent per annum (Table 4).
  - Role of institutions and human capital:
    - Institutional quality and human capital positively associated with growth rate of product quality.
    - A one standard deviation increase in institutional quality associated with a 0.3 standard deviations increase in pace of quality upgrading.
    - A one standard deviation increase in human capital associated with a 0.2 standard deviations increase in pace of quality upgrading.
  - Demand-side constraints:
    - Lower-income countries tend to serve markets that import lower-quality products (Figure 12), but differences are not large enough to constitute a binding constraint on upgrading.
    - On average, poorer exporters have a larger gap between their export quality and the average quality of their trade partners’ imports; slow-converging countries have substantially lower export quality than partners’ average imports.
    - Policy implication: focus on creating a domestic environment conducive to quality upgrading rather than primarily on lowering barriers to entry in destination markets.

### V. Conclusions and policy implications
- Dataset and methodological recap:
  - New dataset covers 178 countries over 1962–2010 with more than 20 million quality estimates.
  - Estimates based on sector-specific quality-augmented gravity that control for production-cost differences and selection bias from distance.
- Main empirical conclusions:
  - Average country-level quality strongly correlated with income per capita.
  - Quality upgrading rapid at early development stages until GDP per capita ≈ $10,000; convergence largely complete by $20,000.
  - Faster growth in export quality associated with faster GDP per capita growth in those countries/time periods.
  - Substantial cross-country and regional heterogeneity in pace of quality upgrading.
  - Manufacturing shows faster convergence than average; agriculture holds substantial upgrading potential for LICs.
- Policy recommendations emphasized in the analysis:
  - Horizontal policies over sector-specific policies:
    - Improve institutional quality and human capital to accelerate quality upgrading.
    - Emphasize general domestic environment conducive to upgrading rather than highly targeted sectoral interventions.
  - Facilitate horizontal diversification:
    - For countries at or near the frontier in existing sectors (e.g., electronics), promote horizontal diversification into sectors with long “quality ladders” to sustain future upgrading opportunities.
    - For countries near bottom of ladders in primary/agricultural sectors, promote within-sector upgrading where feasible and diversify into manufactures where appropriate.
  - Consider agriculture as a source of upgrading:
    - LICs with large agricultural labor shares can achieve substantial quality gains in agriculture and processing.
  - Reducing barriers to entry into new sectors can boost growth by expanding future quality-upgrading possibilities; however, lifting domestic constraints and building capabilities is a priority given limited evidence that destination-market demand is the main constraint.
- Suggested interpretation for policy design:
  - Because product-level quality converges to frontier values and successful export outcomes are often unpredictable, policies that broadly raise capabilities (institutions, human capital) and lower barriers to diversification are likely to be more effective than narrowly targeted sectoral policies.

### VI. Key tables and numeric findings (exact figures preserved from source)
- Dataset and estimation counts:
  - 178 countries; 1962–2010; 45.3 million observations at SITC 4-digit (Revision 1) level; 851 products; more than 20 million quality estimates.
- Table 2. Quality Upgrading: Summary Statistics for Data
  - Growth Rate of Quality 0.00121 0.03703 1.95781 -3.60544
  - Ln Initial Quality -0.22553 0.12213 0.00097 -0.63778
  - Ln Initial GDP per capita 6.86541 1.17895 10.79187 4.64996
  - Institutional Quality 2.03812 6.836427 7.00000 -45.08594
  - Human Capital 18.45108 12.912746 64.15278 0.53906
- Table 3. Quality Upgrading: Panel Regressions, for All Products (selected coefficients and memos)
  - Ln Initial Quality: -0.049 *** (4.2e-4) ; -0.049 *** (4.4e-4) ; -0.046 *** (5.0e-4) ; -0.046 *** (5.1e-4)
  - Ln Initial GDP per capita: 1.5e-4 (1.9e-4) ; 1.1e-4 (2.3e-4) ; -1.5e-4 (2.3e-4)
  - Initial Institutional Quality: 3.8e-5 ** ; 5.2e-5 *** (1.6e-5) (1.7e-5)
  - Initial Human Capital: 2.9e-5 *** (7.2e-6)
  - Memo Items:
    - Observations 237,689 ; 224,139 ; 191,704 ; 181,877
    - Countries 174 ; 173 ; 137 ; 127
    - Products 851 ; 851 ; 851 ; 851
    - Time Period 1962–2008 ; 1962–2008 ; 1965–2008 ; 1965–2008
    - R2 0.070 ; 0.068 ; 0.062 ; 0.063
- Table 4. Quality Upgrading: Panel Regressions, for Manufacturing Alone (selected coefficients and memos)
  - Ln Initial Quality: -0.071 *** (5.9e-4) ; -0.072 *** (6.2e-4) ; -0.066 *** (6.8e-4) ; -0.065 *** (7.0e-4)
  - Ln Initial GDP per capita: 8.9e-4 *** (2.1e-4) ; 8.9e-4 *** (2.4e-4) ; 7.6e-4 *** (2.5e-4)
  - Institutional Quality: 3.3e-5 * (1.7e-5) ; 4.2e-5 ** (1.7 e-5)
  - Human Capital: 1.7e-5 ** (7.7e-6)
  - Memo Items:
    - Observations 156,260 ; 147,907 ; 127,561 ; 121,029
    - Countries 174 ; 173 ; 137 ; 127
    - Products 568 ; 568 ; 568 ; 568
    - Time Period 19623–2008 ; 19623–2008 ; 1965–2008 ; 1965–2008
    - R2 0.103 ; 0.104 ; 0.089 ; 0.089
- Table 1. Selected gravity and interaction results (percent of SITC 4-digit-plus sectors presented in table format in source):
  - Common preferential trade agreement: 82963 0.45 0.38
  - Colonial relationship: 801163 0.43 0.79
  - Common colonizer: 50201614 0.20 0.29
  - Common language: 711495 0.28 0.53
  - Common border: 82963 0.38 0.33
  - Ln (distance): 681076 -1.02 -1.04
  - Ln (distance) * Ln (importer GDP per capita): 61141016 0.04 -0.02
  - Ln (exporter GDP per capita)* Ln (importer GDP per capita): 90542 0.10 0.08
  - Ln (unit value) * Ln (importer GDP per capita): 2388243893 -0.01 0.19

*Source: _wp13108 - References.*

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

### _wp13108 - References

### I. Introduction: scope and contributions
- Dataset coverage and scope:
  - Covers 178 countries over 1962–2010.
  - Provides breakdowns at the SITC 4-digit and BEC 3-digit levels.
  - More than 20 million quality estimates produced; trade dataset contains 45.3 million observations at SITC 4-digit (Revision 1) level.
  - Estimation conducted for 851 products (SITC 4-digit-plus).
- Main contributions:
  - New, extensive estimates of export quality suitable for broad cross-country and long-time analysis.
  - Stylized facts on export quality evolution, sectoral patterns, and the relationship between quality and income.
  - Initial econometric analysis of determinants of quality upgrading.

### II. Estimating product quality: methodology and data
- Core identification and estimation strategy:
  - Starts from unit values p_mxt and relates ln p_mxt to unobservable quality θ_mxt, exporter income per capita y_xt, and distance Dist_mx (equation (1) in text).
  - Specifies a quality-augmented gravity equation separately for each product (equation (2)).
  - Substitutes observable components and estimates via two-stage least squares to address endogeneity of lagged price*importer-income interaction (equation (3)).
  - Instrument: lagged value of p_mxt * ln y_it (up to 5 years back); where unit value in preceding year unavailable, use nearest preceding year up to 5 years; if none, observation excluded.
- Key dataset construction and cleaning rules:
  - Starts with COMTRADE; supplements importer-reported data with exporter-reported data where importer reports missing; U.S. export flows take precedence over importer-reported flows.
  - Uses SITC 4-digit-plus products when quantities reported in multiple unit sets; total products = 851.
  - Exclusion rules for smallness or outliers:
    - Drop SITC 4-digit-plus if: (i) comprises less than 1 percent of total observations or trade value of corresponding SITC 4-digit product; or (ii) has less than 1000 observations and comprises less than 25 percent of total observations or trade value of corresponding SITC 4-digit product.
    - Exclude observations with: (i) quantity of 1; or (ii) total trade value of less than $7,500 at 1989 prices; or (iii) a unit value above the 95th or below the 5th percentile in 1989 prices within a given product.
  - Gravity controls from CEPII include: common border, common language, preferential trade agreement indicator, colonial relationship, and common colonizer; distance is great-circle distance between importer and exporter.
  - Income per capita data from Penn World Tables, version 7.1; preferential trade agreements from WTO RTA database.
- Quality calculation and aggregation:
  - Quality θ_mxt computed by adjusting unit values for production-cost differences and selection bias from distance (equation (4)).
  - Quality and importers’ taste for quality δ not separately identified.
  - Quality estimates normalized by their 90th percentile within each product-year; resulting values typically range between 0 and 1.2.
  - Aggregations performed using current trade values as weights to SITC 4-, 3-, 2-, and 1-digit levels and to country totals; normalization to the 90th percentile repeated at each aggregation step.
  - Also produces aggregations based on BEC and three broad sectors (agriculture, non-agricultural commodities, and manufactures).

### III. Stylized facts on export quality
- Quality versus unit values:
  - Unit values are much more dispersed than quality (Figure 1).
  - Quality and unit values are correlated at lower quality levels, but once quality reaches about 85 percent of the world frontier value, quality and unit values are no longer correlated.
  - Quality evolves gradually: changes of more than 20 percent are rare in any subperiod (1962–80; 1980–95; 1995–2010).
  - Quality increases often not accompanied by unit value increases; some countries register considerable quality increases with stable unit values (e.g., computer and electronics sectors).
- Sectoral examples:
  - Passenger motor cars (SITC 7321):
    - U.S. export quality slightly below world frontier on average; fluctuations led to market share decline when quality fell in the 1980s.
    - Germany: high quality and high prices throughout.
    - Japan: strong quality upgrading through sample period; surpassed German quality in 2000s with moderate price rises → increased market share.
    - Korea: low quality until mid–1980s; significant upgrading in early 1990s and 2000s; prices remained relatively low → market share increased.
  - Apparel (SITC 84):
    - China: quality increased from 70 to 90 percent of world frontier since 1980; market share rose drastically; prices remained low at 40 percent of world frontier.
    - Bangladesh and India: strong market share increases; weaker quality increases → no price increases.
    - Italy: maintained world frontier quality but market share declined as prices rose.
    - Korea and Thailand: past increases in market share with rising quality and stable prices; later diversified away from textiles and retained higher-quality segments with falling market shares.
- Quality ladders and country positions:
  - Lengths of sectoral quality ladders vary considerably across sectors and countries (Figure 5).
  - Examples:
    - Tanzania: concentrated in primary and agricultural exports, near bottom of ladder → large potential for quality upgrading; horizontal diversification to manufactures may help.
    - Vietnam: considerable upgrading in miscellaneous manufactures but still room for further upgrading.
    - Malaysia and China: near world frontier in electronics → limited scope for further upgrading within existing exports; may require horizontal diversification or moving to more sophisticated tasks.
- Quality along the development path:
  - Income per capita correlated with export quality across aggregate, manufacturing, agriculture, and non-agricultural commodities (Figure 6).
  - Speed and thresholds:
    - Quality upgrading particularly rapid until GDP per capita reaches $10,000.
    - Convergence continues at a diminishing rate and is largely complete by GDP per capita of $20,000.
    - Among high-income countries, average export quality levels vary within a narrow band.
  - Agriculture and commodities:
    - Quality can and does increase substantially in agricultural products and commodities; this may reflect shifts to more processed products.
    - Quality ladders vary across subsectors in agriculture and manufacturing (Figure 7).
  - Robustness & subsamples:
    - Within-country changes: quality increases with income even controlling for country fixed effects.
    - Small states (population <1.5 million in 2010 or 2011): quality rises with income particularly sharply for income levels below $10,000.
    - Commodity exporters (commodities >50 percent of exports): potential for quality upgrading exists but may be constrained by exogenous factors like mineral grade.
  - Unit values increase with income at a relatively constant rate, especially for manufacturing.
- Regional and income-group patterns:
  - MICs: steady convergence toward world quality frontier in manufacturing since the 1970s; agriculture convergence started later, mid–1990s.
  - LICs: manufacturing quality improvement started in mid–1990s; agricultural quality not yet taken off; non-agricultural commodity quality has deteriorated since the 1980s.
  - East Asia: particularly fast quality upgrading, especially in manufactures; commodity quality increased in 1970s–1980s due to elementary processing; agriculture improved later (late 1980s).
  - Sub-Saharan Africa and South Asia: lagging but recent signs of convergence; manufacturing quality up sharply in sub-Saharan Africa since late 1990s and in South Asia since early 2000s; agriculture in South Asia may be taking off.
  - MENA: manufacturing quality increased 1960s–1980s then stagnated; agriculture shows no sustained increases.
  - Latin America: export quality stagnated for several decades.
  - Significant cross-country heterogeneity within regions (Figure 11); examples of fast convergers: Japan, Korea, China, Vietnam; slower or stalled convergence: Malaysia, Thailand, India, Indonesia, Bangladesh; in Africa: Uganda, Nigeria, Tanzania, South Africa improved rapidly since mid-1990s; Ghana, Kenya, Cameroon stagnant.

### IV. Determinants of quality upgrading: panel regression findings
- Empirical specification:
  - Growth rate of product quality regressed on initial product quality, initial income per capita, initial institutional quality, initial human capital, and other controls (equation (5)).
  - Variables averaged over 10-year non-overlapping periods.
  - Institutional quality measured by “constraints on the executive variable” from Polity IV (robust to Kaufmann-Kraay-Mastruzzi indicators).
  - Human capital measured by secondary-school completion rate (World Development Indicators).
- Key econometric findings:
  - Convergence:
    - Growth rate of product quality depends negatively on its initial level — unconditional and conditional convergence.
    - Speed of convergence for individual products: about 5 percent per annum (Table 3).
    - Within manufacturing: speed of convergence about 7 percent per annum (Table 4).
  - Role of institutions and human capital:
    - Institutional quality and human capital positively associated with growth rate of product quality.
    - A one standard deviation increase in institutional quality associated with a 0.3 standard deviations increase in pace of quality upgrading.
    - A one standard deviation increase in human capital associated with a 0.2 standard deviations increase in pace of quality upgrading.
  - Demand-side constraints:
    - Lower-income countries tend to serve markets that import lower-quality products (Figure 12), but differences are not large enough to constitute a binding constraint on upgrading.
    - On average, poorer exporters have a larger gap between their export quality and the average quality of their trade partners’ imports; slow-converging countries have substantially lower export quality than partners’ average imports.
    - Policy implication: focus on creating a domestic environment conducive to quality upgrading rather than primarily on lowering barriers to entry in destination markets.

### V. Conclusions and policy implications
- Dataset and methodological recap:
  - New dataset covers 178 countries over 1962–2010 with more than 20 million quality estimates.
  - Estimates based on sector-specific quality-augmented gravity that control for production-cost differences and selection bias from distance.
- Main empirical conclusions:
  - Average country-level quality strongly correlated with income per capita.
  - Quality upgrading rapid at early development stages until GDP per capita ≈ $10,000; convergence largely complete by $20,000.
  - Faster growth in export quality associated with faster GDP per capita growth in those countries/time periods.
  - Substantial cross-country and regional heterogeneity in pace of quality upgrading.
  - Manufacturing shows faster convergence than average; agriculture holds substantial upgrading potential for LICs.
- Policy recommendations emphasized in the analysis:
  - Horizontal policies over sector-specific policies:
    - Improve institutional quality and human capital to accelerate quality upgrading.
    - Emphasize general domestic environment conducive to upgrading rather than highly targeted sectoral interventions.
  - Facilitate horizontal diversification:
    - For countries at or near the frontier in existing sectors (e.g., electronics), promote horizontal diversification into sectors with long “quality ladders” to sustain future upgrading opportunities.
    - For countries near bottom of ladders in primary/agricultural sectors, promote within-sector upgrading where feasible and diversify into manufactures where appropriate.
  - Consider agriculture as a source of upgrading:
    - LICs with large agricultural labor shares can achieve substantial quality gains in agriculture and processing.
  - Reducing barriers to entry into new sectors can boost growth by expanding future quality-upgrading possibilities; however, lifting domestic constraints and building capabilities is a priority given limited evidence that destination-market demand is the main constraint.
- Suggested interpretation for policy design:
  - Because product-level quality converges to frontier values and successful export outcomes are often unpredictable, policies that broadly raise capabilities (institutions, human capital) and lower barriers to diversification are likely to be more effective than narrowly targeted sectoral policies.

*Source: _wp13108 - References.*

### REFERENCES

### _wp13108 - REFERENCES

### References cited
- Asmundson, I., forthcoming, “More World Trade Flows: An Updated Dataset,” (Washington: International Monetary Fund).
- Baldwin, R., and F. Robert-Nicoud, 2010, “Trade-in-goods and Trade-in-tasks: An Integrating Framework,” CEPR Discussion Paper No. 7775.
- Cadot, O., C. Carrere, and V. Strauss-Kahn, 2011, “Export Diversification: What‘s Behind the Hump?” Review of Economics and Statistic, Vol. 93, pp. 590–605.
- Easterly, W., A. Reshef, and J. Schwenkenberg, 2009, “The Power of Exports,” World Bank Policy Research Working Paper No. 5081.
- Feenstra, R., and J. Romalis, 2012, “International Prices and Endogenous Quality,” NBER Working Paper No. 18314.
- Grossman, G., and E. Rossi-Hansberg, 2008, “Trading Tasks: A Simple Theory of Offshoring,” American Economic Review, Vol. 98, pp. 1978–97.
- Hallak, J. C., 2006, “Product quality and the direction of trade,” Journal of International Economics, Vol. 68, pp. 238–65.
- Hallak, J. C., and P. Schott, 2011, “Estimating Cross-country Differences in Product Quality,” Quarterly Journal of Economics, Vol. 126, pp. 417–74.
- Harrigan, J., X. Ma, and V. Shlychkov, 2011, “Export Prices of U.S. Firms,” NBER Working Paper No. 17706.
- Hausmann, R., J. Hwang, and D. Rodrik, 2007, “What You Export Matters,” Journal of Economic Growth, Vol. 12, pp. 1–25.
- Head, K., and T. Mayer, 2013, “Gravity Equations: Toolkit, Cookbook, Workhorse,” in Handbook of International Economics, ed. by Gopinath, Helpman, and Rogoff.
- Heston, A., R. Summers, and B. Aten, 2012, Penn World Table Version 7.1, Center for International Comparisons of Production, Income and Prices at the University of Pennsylvania.
- Hummels, D., and P. Klenow, 2005, “The Variety and Quality of a Nation’s Exports,” American Economic Review, Vol. 95, pp. 704–723.
- Hwang, J., 2007, “Introduction of New Goods, Convergence and Growth,” Job Market Paper, Department of Economics, Harvard University.
- Imbs, J., and R. Wacziarg, 2003, “Stages of Diversification,” American Economic Review, Vol. 93, pp. 63–86.
- Khandelwal, A., 2010, “The Long and Short (of) Quality Ladders,” Review of Economic Studies, Vol. 77, pp. 1450–1476.
- Klinger, B., and D. Lederman, 2004, “Discovery and Development: an Empirical Exploration of “New” Products,” World Bank Policy Research Working Paper Series No. 3450.
- Klinger, B., and D. Lederman, 2006, “Diversification, Innovation, and Imitation Inside the Global Technological Frontier,” World Bank Policy Research Working Paper No. 3872.
- Lederman, D., and W. Maloney, 2012, Does What You Export Matter? In Search of Empirical Guidance for Industrial Policies (Washington: World Bank).
- Mukerji, P., and A. Panagariya. 2009, “Within- and Across-Product Specialization Revisited,” Columbia University, New York.
- Schott, P., 2004, “Across-product Versus Within-product Specialization in International Trade,” Quarterly Journal of Economics, Vol. 119, pp. 646–677.
- Sutton, J., and D. Trefler, 2011, “Deductions from the Export Basket: Capabilities, Wealth and Trade,” NBER Working Paper No. 16834.
- World Development Indicators (Washington: World Bank).
- World Trade Organization, Regional Trade Agreements Database (Geneva, Switzerland: WTO).

### Key findings and methodological notes from tables
- Table 1. Imports: Quality-Augmented Gravity Equations
  - In percent of SITC 4-digit-plus sectors (columns headings and cells presented in table):
    - Positive Coefficients / Negative Coefficients / Median coefficient value / Significant / Insignificant / Significant / Insignificant / This paper / Hallak (2006)
    - Common preferential trade agreement: 82963 0.45 0.38
    - Colonial relationship: 801163 0.43 0.79
    - Common colonizer: 50201614 0.20 0.29
    - Common language: 711495 0.28 0.53
    - Common border: 82963 0.38 0.33
    - Ln (distance): 681076 -1.02 -1.04
    - Ln (distance) * Ln (importer GDP per capita): 61141016 0.04 -0.02
    - Ln (exporter GDP per capita)* Ln (importer GDP per capita): 90542 0.10 0.08
    - Ln (unit value) * Ln (importer GDP per capita): 2388243893 -0.01 0.19
  - Notes: All equations estimated using two stage least squares.

- Table 2. Quality Upgrading: Summary Statistics for Data
  - Variables and summary statistics (preserve formatting and numeric precision exactly):
    - Growth Rate of Quality 0.00121 0.03703 1.95781 -3.60544
    - Ln Initial Quality -0.22553 0.12213 0.00097 -0.63778
    - Ln Initial GDP per capita 6.86541 1.17895 10.79187 4.64996
    - Institutional Quality 2.03812 6.836427 7.00000 -45.08594
    - Human Capital 18.45108 12.912746 64.15278 0.53906
  - Notes: The annualized growth rate of (product) quality is expressed in annualized natural units. Institutional quality is proxied by the “constraints on the executive” variable, from the Polity IV dataset. Human capital is proxied by the secondary-school completion rate.

- Table 3. Quality Upgrading: Panel Regressions, for All Products
  - Dependent variable: Growth in Product Quality
  - Coefficients and standard errors (exact values and significance markers preserved):
    - Ln Initial Quality: -0.049 *** (4.2e-4) ; -0.049 *** (4.4e-4) ; -0.046 *** (5.0e-4) ; -0.046 *** (5.1e-4)
    - Ln Initial GDP per capita: 1.5e-4 (1.9e-4) ; 1.1e-4 (2.3e-4) ; -1.5e-4 (2.3e-4) [positioned in columns per table layout]
    - Initial Institutional Quality: 3.8e-5 ** ; 5.2e-5 *** (1.6e-5) (1.7e-5)
    - Initial Human Capital: 2.9e-5 *** (7.2e-6)
  - Memo Items:
    - Observations 237,689 ; 224,139 ; 191,704 ; 181,877
    - Countries 174 ; 173 ; 137 ; 127
    - Products 851 ; 851 ; 851 ; 851
    - Time Period 1962–2008 ; 1962–2008 ; 1965–2008 ; 1965–2008
    - R2 0.070 ; 0.068 ; 0.062 ; 0.063
  - Notes: All equations estimated using observations averaged over 10-year non-overlapping periods. The dependent variable is the annualized growth rate of product quality. Institutional quality is proxied by the “constraints on the executive” variable, from the Polity IV dataset. Human capital is proxied by the secondary-school completion rate. Other controls include a full set of country, product, and time-period fixed effects. *, **, and *** denote statistical significance at the, respectively, 10 percent, 5 percent, and 1 percent level.

- Table 4. Quality Upgrading: Panel Regressions, for Manufacturing Alone
  - Dependent variable: Growth in Product Quality
  - Coefficients and standard errors (exact values and significance markers preserved):
    - Ln Initial Quality: -0.071 *** (5.9e-4) ; -0.072 *** (6.2e-4) ; -0.066 *** (6.8e-4) ; -0.065 *** (7.0e-4)
    - Ln Initial GDP per capita: 8.9e-4 *** (2.1e-4) ; 8.9e-4 *** (2.4e-4) ; 7.6e-4 *** (2.5e-4)
    - Institutional Quality: 3.3e-5 * (1.7e-5) ; 4.2e-5 ** (1.7 e-5)
    - Human Capital: 1.7e-5 ** (7.7e-6)
  - Memo Items:
    - Observations 156,260 ; 147,907 ; 127,561 ; 121,029
    - Countries 174 ; 173 ; 137 ; 127
    - Products 568 ; 568 ; 568 ; 568
    - Time Period 19623–2008 ; 19623–2008 ; 1965–2008 ; 1965–2008
    - R2 0.103 ; 0.104 ; 0.089 ; 0.089
  - Notes: All equations estimated using observations averaged over 10-year non-overlapping periods. The dependent variable is the annualized growth rate of product quality. Institutional quality is proxied by the “constraints on the executive” variable, from the Polity IV dataset. Human capital is proxied by the secondary-school completion rate. Other controls include a full set of country, product, and time-period fixed effects. *, **, and *** denote statistical significance at the, respectively, 10 percent, 5 percent, and 1 percent level.

### Figure notes and descriptive findings
- Figure 1. Quality and Unit Values
  - Notes: Each dot depicts an exporter-year combination. The 90th percentile is set to unity for both unit values and quality observations.
  - Axes and labels as in source: Quality and Unit value scales with Lowess fit displayed for All sectors, Manufacturing, Agriculture, Non-ag. commodities.

- Figure 2. Changes in Quality and Changes in Unit Values
  - Notes: Each dot depicts one exporter.
  - Panels compare changes in Quality vs Change in Unit Value across periods and income groups:
    - Changes 1962-1980 in Quality vs Unit Values (All Sectors; Manufacturing)
    - Changes 1980-1995 in Quality vs Unit Values (All Sectors; Manufacturing)
    - Changes 1995-2009 in Quality vs Unit Values (All Sectors; Manufacturing)
  - Axes and legend denote High Income, Middle Income, Low Income groupings.

- Figure 3. Quality and Unit Values for Passenger Motor Cars Exports (SITC 7321)
  - Panels for United States, Germany, Japan, Korea showing:
    - Unit Value (90th percentile=1), Quality (90th percentile=1), Share of World Exports over years (1960–2010 or 1970–2010 per panel).
    - Example axis tick values in source: Share of World Exports .05 .1 .15 .2 .25 .3; Unit Value .4 .5 .6 .7 .8 .9 1 1.1; Quality .75 .85 .95 1.05.

- Figures 4–12 (captions and high-level descriptors)
  - Figure 4. Quality and Unit Values for Apparel Exports (SITC 84).
  - Figure 5. Quality Ladders.
  - Figure 6. Quality, Unit Values, and GDP per capita.
  - Figure 7. Quality in Agriculture and Manufacturing.
  - Figure 8. Quality, Unit Values, and GDP per capita: Within-Country Variation.
    - Notes: Figures for small states and commodity exporters use both within- and cross-country variation.
  - Figure 9. Export Quality by Income Group over Time.
  - Figure 10. Quality Upgrading by Region.
  - Figure 11. Country-level Heterogeneity in Quality Upgrading in Asia and Africa.
  - Figure 12. Quality Upgrading and Destination Markets.
    - Visual elements noted in source: categories for in Africa / in Asia and income groups (Low income exporters, Lower middle income exporters, Upper middle income exporters, High income exporters), and labels such as Average quality demanded in destination markets; Quality exported; Export quality relative to destination markets in 2009 (World Frontier=1); "Slow convergers:" fragment present in source.

*Source: _wp13108 - REFERENCES*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13108.pdf_