## _030514a ## Source details **Canonical URL:** [_030514a](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2014/_030514a.pdf) ## Other formats - [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2014/_030514a.pdf.md) - [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2014/_030514a.pdf.json) --- ### Key findings on diversification and growth - Export diversification (LICs, nonlinear IVBMA/2SLS results): - A one standard deviation increase in export diversification raises the average annual growth rate by 0.8 percentage points for LICs. - 2SLS specification reported magnitude: about 0.9 percentage points for LICs (100x(-.009)x(-1.029) = 0.926%). - Output (real) diversification (IVBMA results): - A one standard deviation increase in output diversification in LICs raises average annual growth by about 1.4 percentage points. - Standard deviation of output diversification in low income countries: 0.078. - Composite coefficient for Diversification*LowIncome: -0.176 (footnote calculation yields 100x(-0.078)x(-0.176)=1.373 percentage points referenced). - Other robust growth determinants consistently identified: - Initial GDP, Population Growth, Investment, Governance Quality, and Government Expenditures. ### Data, sample, and measurement - Sample and periods: - Sample covers 1965 to 2009. - Dataset includes 84 countries using non-overlapping five-year periods. - Country-period observations: 583 (export diversification regressions), 531 (output diversification regressions). - Exclusions and definitions: - Resource-rich economies excluded if resource rents > 20 percent of GDP. - Small-state threshold: population less than 1.5 million (only Sierra Leone in regression sample; reported robustness without it). - Dependent variable and main measures: - Dependent variable: average growth rate of GDP per capita during each five-year period (Per capita income source: Penn World Tables 7.1). - Primary export diversification measure: Total Theil index; decomposed into extensive (between Theil) and intensive (within Theil) margins (lower Theil = higher diversification). - Output diversification measure: Total Theil index of value-added shares in seven real subsectors (UN sectoral database). ### Econometric approach and identification - Empirical strategy: - Extend Durlauf et al. (2008) with longer panel data and trade-diversification measures. - Address model uncertainty and endogeneity using Instrumental Variable Bayesian Model Averaging (IVBMA). - IVBMA features: - Averages coefficients over empirical models weighted by model quality. - Produces posterior inclusion probabilities (PIP) and conditional means and standard deviations for coefficients. - PIP interpretation thresholds (Jeffreys; Kass and Raftery) reported: - PIP < 50 percent: evidence against an effect. - 50–75 percent: weak evidence. - 75–95 percent: positive evidence. - 95–99 percent: strong evidence. - > 99 percent: decisive evidence. - A regressor considered “effective” if PIP > 50 percent. - IVBMA diagnostics and validity: - IVBMA-Sargan test indicates instrument validity in reported IVBMA specifications (Sargan test p-value reported as 1.000 in Appendix Table A2 where shown). - Appendix reporting includes PIP, conditional mean, and conditional SD for many regressors (selected reported values reproduced in source tables). ### Instrumentation strategy (export diversification) - Instruments used to address simultaneity (Frankel and Romer spirit): - log of a country’s land area (lLand). - dummy for landlocked countries (Landlocked). - log of a country’s population (lPop). - Rationale: geographic features are predetermined with respect to growth and drive export diversification (population supports product variety; land area supports specialization clusters; landlocked status reduces trade). ### Empirical results — export diversification and volatility - Export diversification reduces growth volatility in developing countries: - System GMM (developing countries, 1962–2010) reports positive coefficients on Theil and HFI measures (lower Theil/HFI = higher diversification implies reduced volatility). - Interaction term evidence: diversification reduces growth volatility more when trade openness is higher (examples: Within Theil x Openness = 0.0515** for developing countries; Theil x Openness = 0.0237** for all countries). - Stylized facts and spurts: - 61 diversification spurts in 51 developing countries post-1962. - Spurts lasted on average 13 years; in East Asia and Pacific spurts averaged 20 years. - In LICs, growth volatility decreases by 1.5 percentage points in the wake of diversification spurts. ### Empirical results — margins of diversification - Decomposition (export Theil into extensive and intensive margins): - For LICs, both extensive (new products) and intensive (more balanced shares) margins contribute to growth gains. - Some evidence lower- and upper-middle-income countries may benefit more via the extensive margin. ### Sectoral productivity, structural transformation, and policy drivers - Sectoral patterns and productivity gaps: - Average LIC manufacturing share remains about 12 percent; agricultural employment remains over 60 percent in average SSA country despite low agricultural value added. - Agricultural productivity in LICs averaged about 30 percent of non-agricultural labor productivity in 2005. - Agriculture in SSA often less than five percent of U.S. level in labor productivity. - Drivers of within-sector productivity and sectoral shifts (panel GMM, 28 LICs, 1995–2010): - Agriculture productivity positively associated with trade liberalization (tariff removal) and removal of interest rate controls (coefficient examples: trade liberalization on agriculture 0.175 [0.093]*; removal of interest rate control on agriculture 0.640 [0.275]**). - Manufacturing productivity linked to domestic financial sector reform, FDI liberalization (0.369 [0.214]*), road infrastructure (0.091 [0.028]***), and tertiary education (0.017 [0.009]*). - Services productivity associated with networks reform (telecommunications) (0.138 [0.063]**). - Sectoral shifts: - Trade liberalization associated with reallocation from agriculture to manufacturing and services. - Infrastructure, electricity liberalization, eased credit, and regulatory reform support industry employment share and labor mobility. - Policy priorities indicated for short-term productivity and structural transformation: - Removing tariff barriers, reforming domestic financial sector (including interest rate controls), liberalizing network industries, liberalizing FDI/capital account, improving road infrastructure, enhancing tertiary education, and easing regulatory burdens. ### Quality upgrading (horizontal vs vertical transformation) - Quality measurement and coverage: - Multi-level export quality database (Henn, Papageorgiou, Spatafora, 2013): 178 countries, 851 products, 1962–2010; quality normalized to world frontier (90th percentile); quality values typically range between 0 and 1.2. - Stylized facts: - Quality upgrading most pronounced in East Asia manufacturing; apparel an important initial sector. - Non-fragile LICs upgraded quality considerably; fragile LICs lag. - Quality increases tend to raise export prices until quality reaches about 80–85 percent of world frontier; beyond that, price gains taper. - Growth links and magnitudes: - In 1995–2010, a 0.1 quality increase in agriculture associated with additional annual GDP per capita growth of 0.5 percentage points (holds for LICs). - A 0.1 quality increase in manufacturing associated with additional annual GDP per capita growth of 1.0 percentage point (strength similar for LICs). - Fast quality convergers grew about one percentage point per year more than slow convergers during 1995–2010. - Determinants of faster quality growth: - Institutional quality, human capital, liberal trade regimes, liberal agricultural policy, and liberalized domestic financial markets. - Empirical associations: one standard deviation increase in institutional quality → 0.3 standard deviations faster quality upgrading; one standard deviation increase in human capital → 0.2 standard deviations faster quality upgrading (Henn et al., 2013 referenced). ### Case study summaries and country lessons - Vietnam (Đổi Mới since 1986): - Reforms in agriculture, prices, external sector, and macro stabilization; WTO accession in 2007. - FDI and export transformation from oil to light manufacturing (textiles, footwear, electronics). - Policy lesson: successive reform waves and broad liberalization spurred diversification; further gains require higher education and less administrative steering. - Tanzania: - Reforms mid-1980s onward; growth accelerated to average 6¾ percent during 2000–2011. - Gold rose to be single largest export (9¼ percent of GDP in 2011; 30 percent of GNFS). - Manufacturing exports rose; economy moved from least to most diversified in sample by 2011. - Policy needs: sustain growth, raise manufacturing share, invest in infrastructure and human capital; prepare for natural gas discovery risks with resource-management framework. - Angola: - Post-2001 reconstruction financed by oil; annual GDP growth 2002–08 averaged 15¼ percent; oil accounts for 99 percent of exports in 2011. - Agriculture underdeveloped despite fertile land; diversification constrained by oil dominance and governance challenges. - Policy trade-offs: broad-based reforms, infrastructure, human capital, and governance to avoid Dutch-disease effects and rent-seeking. ### Summary of key numeric values and diagnostics - Countries in sample: 84. - Sample period: 1965–2009 (export analyses 1962–2010 referenced for volatility; output diversification 1970–2010 where value-added data start). - Country-period observations: 583 (export diversification regressions), 531 (output diversification regressions). - Resource-rich exclusion threshold: > 20 percent of GDP from resource rents. - Small-state population threshold: < 1.5 million. - Export diversification impact (LICs, nonlinear IVBMA): ~0.8 percentage points per one standard deviation increase. - Output diversification impact (LICs, nonlinear IVBMA): ~1.4 percentage points per one standard deviation increase. - Output diversification standard deviation (LICs): 0.078. - Composite coefficient Diversification*LowIncome: -0.176. - Selected IVBMA PIP and coefficient examples (reported in source tables): - Initial GDP: PIP 1, Cond. Mean -0.015, Cond. SD 0.002. - GovernanceQuality: PIP 0.955, Cond. Mean 0.009, Cond. SD 0.003. - Investment: PIP 0.999, Cond. Mean 0.013, Cond. SD 0.003. - Diversification (global): PIP 0.283, Cond. Mean -0.004, Cond. SD 0.003. - Diversification Low Income (composite): PIP 0.997, Cond. Mean -0.008, Cond. SD 0.003. - Diversification (Total Theil, output): PIP 0.769, Cond. Mean -0.117, Cond. SD 0.065. - Diversification Low Income (output composite): PIP 0.988, Cond. Mean -0.176, Cond. SD 0.079. - Appendix table diagnostics: - Sargan test p-value reported as 1.000 in Appendix Table A2 (where shown). - Sargan test p-value reported as 0.000 in Appendix Table A3 (indicating rejection in those IV specifications). ### Policy implications and recommendations (synthesized) - Prioritize export and output diversification as growth targets in LICs: - Export diversification estimated to raise LIC growth by ~0.8 percentage points per one standard deviation increase. - Output diversification yields even larger benefits (~1.4 percentage points per one standard deviation increase). - Use policy packages that foster both horizontal diversification and quality upgrading: - Facilitate entry into new products and markets (horizontal diversification) and improve institutional quality and human capital to hasten quality convergence. - Emphasize broad-based reforms over narrow, highly targeted industrial policies where governance is weak; sector-specific support can work but carries risks of rent-seeking. - Address volatility and resilience: - Diversification reduces growth volatility—effects amplified with greater trade openness. - Policies to increase intensive diversification (equalizing export shares) are particularly useful to lower volatility. - Sector-specific policy levers: - Agriculture: remove tariff barriers, reform domestic finance (remove interest rate controls), improve logistics (air transport for perishables). - Manufacturing: deepen financial reform, liberalize FDI, improve road infrastructure, and expand tertiary education. - Services: liberalize network industries, improve road connectivity for distribution services. - Resource management and sequencing: - For resource-rich or potentially resource-rich LICs, design comprehensive frameworks to manage natural resource wealth to avoid Dutch disease and preserve competitiveness of tradables. - Institutional and human capital investment: - Strengthen governance and raise human capital; one standard deviation improvements in these correlate with materially faster quality upgrading and growth dividends. *Source: _030514a - References and Background Notes (IMF), March 5, 2014.* ### References ________________________________________________________________________________________ 15 ### _030514a - References ### Overview and document structure - Document contains: References; Appendix I. OLS and 2SLS Growth Regression; Appendix II. IVBMA Methodology. - Appendix tables listed: A1. Export Diversification; A2. Real Sector Diversification; A3. OLS and 2SLS Estimation; A4. Data Appendix. - Major sections and background notes include: "Diversification and Volatility"; "Diversification and Transformation in LICs—Background Notes"; "Structural Transformation and Sectoral Productivity in Low-Income Countries"; "Opportunities for Quality Upgrading in Low-Income Countries"; "Experiences and Policy Lessons from Five Case Studies". - Figures and tables enumerated (selected): Figures 1–36; Tables 1–7. - Date on document: March 5, 2014. - Prepared by: Theo Eicher (University of Washington), David Kuenzel (University of Washington), and Ke Wang (IMF). ### Key substantive findings (as presented) - Export diversification is identified as a crucial determinant of economic growth for LICs: - A one standard deviation increase in export diversification increases the average annual growth rate by 0.8 percentage points for LICs. - Output diversification matters for growth: - A one standard deviation increase in output diversification in LICs raises the average annual growth rate by about 1.4 percentage points. - Other robust growth determinants identified by the approach include: - initial GDP, population growth, investment, governance quality, and government expenditures. ### Conceptual and theoretical framing - Structural transformation defined as the dynamic reallocation of resources from less productive to more productive sectors (McMillan and Rodrik, 2011; Lin, 2012). - Diversification conceptualized along two margins (Papageorgiou and Spatafora, 2012): - Extensive margin: number of different export sectors. - Intensive margin: diversification of export volumes across active sectors; countries with export revenues driven by a few sectors may appear less diversified on the intensive margin. - Theoretical channels linking diversification and growth: - Reduced growth volatility via portfolio selection effects. - Gradual allocation of resources to more productive uses (Acemoglu and Zillibotti, 1997). - Recognition of simultaneity and potential feedback between growth and diversification; structural transformation examples include East Asian Tigers and ex-Soviet economies. ### Econometric methodology and identification - Empirical approach extends Durlauf et al. (2008) with: - Longer panel data. - Introduction of trade diversification measures. - Use of Instrumental Variable Bayesian Model Averaging (IVBMA) to address model uncertainty and endogeneity. - IVBMA features and rationale: - Averages coefficients over empirical models weighted by model quality. - Produces posterior inclusion probabilities (PIP) to quantify the probability that a regressor affects the dependent variable. - Interpretation thresholds for PIP (Jeffreys; Kass and Raftery): - PIP lower than 50 percent: evidence against an effect. - 50–75 percent: weak evidence for an effect. - 75–95 percent: positive evidence. - 95–99 percent: strong evidence. - higher than 99 percent: decisive evidence. - A regressor is considered “effective” if its PIP exceeds 50 percent. - Comparison with other model-selection and endogeneity-handling approaches: - Contrasts with Extreme Bound Analysis (Leamer, 1978) and standard Bayesian Model Averaging approaches that do not handle endogeneity. - References to methodological extensions: Eicher et al. (2009), Moral-Benito (2012), Chen et al. (2009), Koop et al. (2012), Karl and Lenkoski (2012). ### Instrumentation strategy and identification for export diversification - Instruments used to address simultaneity (in the spirit of Frankel and Romer, 1999): - log of a country’s land area, - a dummy for landlocked countries, - log of a country’s population. - Rationale: - Geographic features are predetermined with respect to growth rates and are important drivers of export diversification (e.g., large population supports more types of products; large land area supports specialization clusters; landlocked status reduces international trade and thus export diversification). ### Policy implications highlighted - Export diversification should be an important growth policy target for LICs given its estimated quantitative impact on average annual growth (0.8 percentage points per one standard deviation increase). - Encouraging output diversification in LICs can yield larger growth benefits (about 1.4 percentage points per one standard deviation increase). *Source: _030514a - References (IMF), March 5, 2014.* ### 14. Using non-overlapping five-year periods, the dataset includes 84 countries from the ### _030514a - 14. Using non-overlapping five-year periods, the dataset includes 84 countries from the ### Data and sample - Sample covers the period 1965 to 2009 and comprises 583 country-period observations. - The dataset includes 84 countries using non-overlapping five-year periods. - Resource-rich economies defined as generating more than 20 percent of their GDP with resource rents (World Development Indicators) are excluded. - Rationale: resource-rich countries are sizable outliers with unusually low export diversification relative to income. - Small states were removed; the only small state (population less than 1.5 million) in the regression sample is Sierra Leone. - Regression results remain at the same significance level and coefficients hardly change without this country. - Dependent variable: average growth rate of GDP per capita during each five-year period. - Per capita income data source: Penn World Tables 7.1. - Primary measure of export diversification: Total Theil index. - Decomposition used into extensive margin (between Theil) and intensive margin (within Theil). - Note: lower values of the Theil index indicate higher diversification. - Output diversification measured by the total Theil index of value-added shares in seven real subsectors (UN sectoral database). ### Covariates, instruments, and theoretical proxies - All additional covariates and instruments are obtained from the growth determinants study of DKT (2008), updated by Henderson et al., (2011). - DKT base variable selection on Barro (2003) and introduce proxies for seven different growth theories: i) Neoclassical growth theory regressors: - initial per capita income; - per-period averages of population growth; - investment to GDP ratio; - education (share of the working population with secondary schooling times the rate of successful completion of secondary school). - These four variables are instrumented with one-period lagged values. ii) Demographic change proxies: - reciprocal of life expectancy at age one; - logarithm of the total fertility rate. - Both assumed exogenous. iii) Macroeconomic policy proxies: - average ratio of government consumption to GDP; - openness (exports + imports) over GDP filtered for land mass and population; - average change in the CPI. - All three instrumented with respective lagged values. iv) Geography proxies: - land area within 100km of an ice-free coast; - percentage of tropical land area. - Assumed exogenous. v) Institutions proxies: - risk of expropriation (lagged values used to instrument current value); - constraints on the executive; - World Bank governance index; - dummy variables for English and French legal origin. - All other institutional variables treated as exogenous. vi) Religion proxies: - shares adhering to Eastern, Hindu, Jewish, Muslim, Orthodox, Protestant, and other religions. - Respective shares in 1900 used as instruments. vii) Fractionalization proxies: - linguistic fractionalization index; - ethnic tension index. - Both assumed exogenous. - Dataset details, variable sources and definitions are described in the appendix of the source. ### Empirical strategy and estimation - Three-step empirical strategy: - Introduce export diversification as a potential growth determinant. - Address model uncertainty due to many competing growth theories and candidate regressors. - Examine importance of controlling for endogeneity in growth regressions. - Estimator: Instrumental Variable Bayesian Model Averaging (IVBMA). - IVBMA results reported with posterior inclusion probabilities (PIP), conditional means and standard deviations for coefficients; coefficients interpretable as in standard OLS and 2SLS. - IVBMA-Sargan test (Eicher et al., 2009) indicates instrument validity in all IVBMA specifications reported. ### Key empirical findings — Export diversification and growth - In the global sample with a linear diversification term (Table 1, column 1): - Traditional growth determinants with highest effect thresholds: Initial GDP, Government Quality, Investment, Population Growth, Government Expenditure; - Followed by Protestant Fraction, Sub Saharan Africa dummy, and Inflation. - Export Diversification does not have an effect on growth volatility in the global sample in this linear specification. - Nonlinear specification (Table 1, column 2) — export diversification for LICs: - Export diversification has a decisive impact on growth for LICs once nonlinearities are introduced. - Using column 2 results: a one standard deviation increase in LICs’ export diversification raises their growth rate by about 0.8 percentage points. - IVBMA-Sargan tests indicate instrument validity in all IVBMA specifications in Table 1. - Conclusion: export diversification is crucial for growth in LICs. - Decomposition into margins (Table 1, columns 3 and 4): - Replacing Total Theil with intensive (within) and extensive (between) Theil indices yields very similar conclusions for LICs. - Extensive Theil indicates diversification by expanding to new products; intensive Theil indicates diversification by equalizing product shares. - LICs can stimulate growth by diversifying exports both at the extensive and intensive margins. - Some evidence that lower- and upper-middle income countries may benefit from increasing diversification at the extensive margin. - Observations in export diversification regressions: 583. ### Key empirical findings — Output diversification and growth - Output diversification (Real Theil) IVBMA results (Table 2): - Linear output diversification specification (column 1): output diversification is significant in the linear specification (contrast with export diversification linear result). - Nonlinear output diversification specification (column 2): LICs are the likely driver of the aggregate effect. - Quantitative magnitude: a one standard deviation increase in output diversification in LICs raises their average annual growth rate by about 1.4 percentage points. - Footnote calculation: standard deviation of output diversification in low income countries is 0.078; composite coefficient for Diversification*LowIncome is -0.176, yielding predicted increase 100x(-0.078)x(-0.176)=1.373 percentage points. - This effect is even greater than the one found for export diversification. - Observations in output diversification regressions: 531. ### Interpretation and policy-relevant implications - Robustness: - Results are robust to endogeneity and model uncertainty through use of IVBMA. - Both export diversification (intensive and extensive margins) and output diversification are identified as drivers of economic growth in LICs. - Policy implications: - Countries at early stages of development could benefit considerably by diversifying their exports and production structures. - At later stages of development, export diversification appears more a by-product of prosperity than a causal driver. - Mechanisms through which early diversification may drive development: - Insurance against idiosyncratic sectoral shocks, particularly where countries produce few goods for export (e.g., agricultural products and natural resources). - Greater export diversification at early stages increases likelihood of moving into new products and spurring further development. - The set of growth determinants identified by IVBMA is parsimonious and consistent with theory: - Supports neoclassical growth model (Initial GDP, Investment, Population Growth) and growth theories emphasizing productive government expenditures and institutional quality (Government Quality, Government Expenditure). ### Summary of key numeric values and counts - 84 countries in sample. - Period: 1965 to 2009. - Country-period observations: 583 (export diversification regressions), 531 (output diversification regressions). - Resource-rich exclusion threshold: more than 20 percent of GDP from resource rents. - Small-state population threshold: less than 1.5 million. - Data source for per capita income: Penn World Tables 7.1. - Estimated impact magnitudes: - Export diversification (LICs, nonlinear): ~0.8 percentage points increase in growth per one standard deviation increase. - Output diversification (LICs, nonlinear): ~1.4 percentage points increase in growth per one standard deviation increase. - Standard deviation of output diversification in LICs: 0.078. - Composite coefficient reported for Diversification*LowIncome: -0.176 (footnote calculation yields 1.373 percentage points). *Source: DIVERSIFICATION AND TRANSFORMATION IN LICs—BACKGROUND NOTES, INTERNATIONAL MONETARY FUND* ### Appendix I. OLS and 2SLS Growth Regression ### Appendix I. OLS and 2SLS Growth Regression ### Overview - This appendix compares standard OLS and 2SLS growth regression results with export diversification (see Table A3 in the source). - The analysis reports OLS baseline results, a specification allowing nonlinearities via income dummies and interactions with export diversification, and 2SLS results that control for endogeneity of trade and other regressors. - IVBMA (Instrumental Variable Bayesian Model Averaging) methodology is presented in Appendix II to address instrument selection and model uncertainty in a two-stage framework. ### OLS Results — baseline and diversification specifications - OLS baseline (column 1) produces results roughly comparable to Barro (2003): - InitialGDP, Investment, and PopulationGrowth are significant consistent with the neoclassical model. - Institutional factors GovExpenditures and ExecutiveConstraints are significant. - One religious measure (Jewish) is significant. - FilteredOpenness is significant at the 10 percent level. - Inflation, Fertility, and LandNearCostPct are also significant. - Linear export diversification term (Diversification) in column 2 is not significant in the global OLS panel. - Nonlinear specification (column 3) introducing World Bank income dummies and interactions with Diversification finds no significant effect of export diversification for any income group in OLS. - Note: absence of OLS effect is consistent with evidence of feedback between growth and trade. ### Controlling for endogeneity — 2SLS findings and diagnostics - Column 4 acknowledges endogeneity for trade and 18 other growth determinants (list of endogenous regressors provided in source). - Identification diagnostics: - Angrist-Pischke first-stage chi-squared and F statistics reject under-identification and weak identification at the 1 percent level for all endogenous variables. - The Sargan-Hansen J statistic rejects instrument validity in the 2SLS regression, indicating instrument validity concerns and suggesting a more parsimonious 2SLS specification may be preferred. - Main 2SLS substantive findings: - 2SLS results largely align with OLS determinants in column 3. - After controlling for endogeneity, export diversification becomes significant for LICs. - Magnitude for LICs: a one standard deviation increase in export diversification raises average annual growth in LICs by about 0.9 percentage points. - Source computation preserved: coefficient of -.009 and the 1.029 standard deviation of export diversification for low income countries imply that a one standard deviation increase in diversification increases growth by about 0.9 percentage points (100x(-.009)x(-1.029) = 0.926%). - Investment and the marginally significant variables ExecutiveConstraint and Fertility lose significance in the 2SLS approach. - The loss of significance for Investment is consistent with previous panel studies once endogeneity is controlled for. ### IVBMA methodology (summary of mechanics) - IVBMA combines IV and BMA methodologies, accounting for model uncertainty in both first and second stages. - First stage: BMA identifies effective instruments; posterior distributions and posterior model probabilities are computed for candidate first-stage models. - Posterior mean in stage 1: iBMAˆθ = sum over models of posterior means weighted by posterior model probabilities. - Posterior variance accounts for within-model variance and variance of estimates across models (equation (5) in source). - Inclusion probability (posterior inclusion probability) equals the sum of posterior probabilities of models that contain the variable (equation (6) in source). - Second stage: predicted values from each first-stage model, wi, generate second-stage model posterior probabilities and estimates. - IVBMA estimate is the weighted sum of averaged posterior IV means obtained using fitted values from each first-stage model, with weights reflecting first-stage model quality. - Posterior variance and inclusion probabilities in IVBMA reflect uncertainty and model weighting across both stages. - Interpretation: underperforming instrument models are deemphasized; strong instrument models receive higher posterior weights. ### Selected IVBMA / Extended DKT conditional posterior results (from Appendix Tables A1–A2) - Table formatting in source reports PIP, Cond. Mean, and Cond. SD for many regressors under multiple specifications (Extended DKT, Extended DKT (Total Theil), Extended DKT (Between Theil), Extended DKT (Within Theil), and IVBMA). Selected entries are reproduced exactly as they appear in the source tables: - Initial GDP: PIP 1, Cond. Mean -0.015, Cond. SD 0.002 (first reported column) - GovernanceQuality: PIP 0.955, Cond. Mean 0.009, Cond. SD 0.003 (first reported column) - Investment: PIP 0.999, Cond. Mean 0.013, Cond. SD 0.003 (first reported column) - GovExpend: PIP 0.856, Cond. Mean -0.079, Cond. SD 0.029 (first reported column) - PopulationGrowth: PIP 0.988, Cond. Mean -0.05, Cond. SD 0.012 (first reported column) - JewishFraction: PIP 0.289, Cond. Mean 0.032, Cond. SD 0.015 (first reported column) - FilteredOpenness: PIP 0.04, Cond. Mean 0.003, Cond. SD 0.004 (first reported column) - Diversification (global): PIP 0.283, Cond. Mean -0.004, Cond. SD 0.003 (first reported column) - Diversification Low Income (composite): PIP 0.997, Cond. Mean -0.008, Cond. SD 0.003 (reported in table) - Diversification Lower Middle Income (composite): PIP 0.217, Cond. Mean -0.001, Cond. SD 0.003 (reported in table) - Diversification Upper Middle Income (composite): PIP 0.174, Cond. Mean -0.001, Cond. SD 0.003 (reported in table) - Appendix Table A2 (Output Diversification, IVBMA Estimates Production Diversification (Total Theil)) — selected entries: - Initial GDP: PIP 0.924, Cond. Mean -0.012, Cond. SD 0.004 (Extended DKT IVBMA); PIP 1.000, Cond. Mean -0.020, Cond. SD 0.003 (Extended DKT IVBMA second column) - GovernanceQuality: PIP 0.995, Cond. Mean 0.014, Cond. SD 0.004; PIP 1.000, Cond. Mean 0.013, Cond. SD 0.003 - Investment: PIP 0.949, Cond. Mean 0.012, Cond. SD 0.003; PIP 1.000, Cond. Mean 0.013, Cond. SD 0.003 - GovExpend: PIP 0.293, Cond. Mean -0.066, Cond. SD 0.046; PIP 0.996, Cond. Mean -0.128, Cond. SD 0.031 - PopulationGrowth: PIP 0.838, Cond. Mean -0.038, Cond. SD 0.014; PIP 0.998, Cond. Mean -0.060, Cond. SD 0.013 - Diversification (Total Theil): PIP 0.769, Cond. Mean -0.117, Cond. SD 0.065 (Extended DKT IVBMA first column); PIP 0.186, Cond. Mean 0.006, Cond. SD 0.039 (second column) - Diversification Low Income (composite): PIP 0.988, Cond. Mean -0.176, Cond. SD 0.079 (reported) - Diversification*LowIncome: PIP 0.983, Cond. Mean -0.178, Cond. SD 0.078 (reported) - Additional table diagnostics reported in source: - Sargan test p-value reported as 1.000 in Appendix Table A2 (where shown). - Observations reported as 583 in Appendix Table A1 and 531 in Appendix Table A2 (as shown in the source tables). *Source: Appendix I and Appendix II, "_030514a - Appendix I. OLS and 2SLS Growth Regression" (IMF background notes). ### Appendix Table A3. OLS and 2SLS Estimation ### Appendix Table A3. OLS and 2SLS Estimation ### Key estimation results (Export Diversification — selected coefficients and statistics) - Initial GDP: -0.010*** 0.003; -0.010*** 0.003; -0.011*** 0.003; -0.018*** 0.004 - GovernanceQuality: 0.004 0.003; 0.004 0.003; 0.005 0.003; 0.009*** 0.003 - Investment: 0.010*** 0.003; 0.010*** 0.003; 0.011*** 0.003; 0.006 0.004 - GovExpend: -0.114*** 0.026; -0.114*** 0.026; -0.115*** 0.027; -0.128*** 0.041 - PopulationGrowth: -0.038*** 0.011; -0.038*** 0.011; -0.040*** 0.011; -0.055** 0.022 - JewishFraction: 0.039*** 0.009; 0.040*** 0.009; 0.038*** 0.010; 0.063*** 0.017 - LegalOriginsUK: 0.004 0.003; 0.004 0.003; 0.007** 0.003; 0.008* 0.005 - Inflation: -0.000*** 0.000; -0.000*** 0.000; -0.000*** 0.000; -0.000** 0.000 - Fertility: -0.003* 0.001; -0.002 0.002; -0.003* 0.002; -0.002 0.002 - LinguisticFractionalization: -0.008 0.005; -0.009 0.005; -0.007 0.006; -0.013** 0.006 - ExecutiveConstraint: -0.006* 0.004; -0.006* 0.004; -0.006* 0.004; -0.001 0.005 - FilteredOpenness: 0.007* 0.004; 0.007* 0.004; 0.005 0.004; 0.005 0.005 - LandNearCoastPct: -0.007* 0.004; -0.007* 0.004; -0.006 0.004; -0.008** 0.004 - Diversification (composite interactions reported): Diversification Low Income♠: -0.001 0.003; -0.009** 0.004 - Diversification Lower Medium Income♠: -0.003 0.002; -0.005 0.006 - Diversification Upper Medium Income♠: 0.002 0.008; -0.006 0.017 - Diversification*LowIncome: -0.003 0.004; -0.009 0.006 - Diversification*MedIncome: -0.004 0.004; -0.005 0.007 - Diversification*UpperMedIncome: 0.000 0.008; -0.006 0.017 - LowIncomeDummy: 0.005 0.013; 0.024 0.019 - LowerMedIncomeDummy: 0.017* 0.009; 0.019 0.014 - UpperMedIncomeDummy: 0.001 0.017; 0.013 0.034 - Sargan test p-value: 0.000 - R-squared: 0.408; 0.409; 0.417; 0.376 - Observations: 583; 583; 583; 583 Note: ♠ Composite coefficient comprised of Diversification and Diversification*CountryIncome interaction, calculated using Delta Method. ### Notable significance indicators reported in the table - Triple-star significance reported for: Initial GDP, Investment, GovExpend, PopulationGrowth, JewishFraction, and Inflation in several specifications. - Sargan test p-value reported as 0.000 (indicating rejection in overidentification test for the IV specifications reported). - Multiple interaction terms with country income groups reported with composite coefficients and standard errors. --- ### Data Appendix — variable summary statistics (selected variables) - Extensive Theil (Exports): Mean 0.413; SD 0.423; Min -0.025; Max 2.667 - Definition: Average extensive Theil measure of export diversification. Source: IMF - Intensive Theil (Exports): Mean 2.489; SD 0.882; Min 0.970; Max 5.751 - Definition: Average intensive Theil measure of export diversification. Source: IMF - Total Theil (Exports): Mean 2.902; SD 1.057; Min 0.992; Max 5.939 - Total Theil (Output): Mean 0.236; SD 0.101; Min 0.054; Max 0.516 - Initial GDP: Mean 8.542; SD 1.087; Min 6.177; Max 10.806 - Definition: Logarithm of initial per capita GDP in each period. - Investment (ratio of investment to GDP): Mean 2.746; SD 0.537; Min 1.097; Max 4.515 - GovernanceQuality: Mean 0.347; SD 0.902; Min -1.690; Max 1.930 - Definition: Average Composite Governance index (average of six governance variables). Source: World Bank - GovExpend (government consumption to GDP): Mean 0.149; SD 0.054; Min 0.041; Max 0.387 - HHI (Herfindahl export concentration): Mean 0.125; SD 0.154; Min 0.002; Max 0.859 - PopulationGrowth: Mean -2.718; SD 0.163; Min -3.201; Max -2.204 - Definition: Logarithm of average population growth rate plus 0.05. - FilteredOpenness: Mean -0.037; SD 0.301; Min -0.505; Max 1.497 - Definition: Average ratio exports plus imports to GDP, filtered for relation to logs of population and area. - Inflation: Mean 12.952; SD 23.312; Min -3.079; Max 270.65 - Definition: Average consumer price inflation rate. - JewishFraction: Mean 0.015; SD 0.103; Min 0.000; Max 0.896 - MuslimFraction: Mean 0.193; SD 0.331; Min 0.000; Max 0.995 - LinguisticFractionalization: Mean 0.350; SD 0.303; Min 0.000; Max 0.923 - EthnicFractionalization: Mean 0.397; SD 0.260; Min 0.002; Max 0.930 - LifeExpectancy (reciprocals at age 1): Mean 0.201; SD 0.492; Min 0.012; Max 2.253 - LandNearCoastPct: Mean 0.508; SD 0.345; Min 0.000; Max 1.000 - LandTropicsPct: Mean 0.307; SD 0.395; Min 0.000; Max 1.000 - LowIncomeDummy: Mean 0.216; SD 0.412; Min 0.000; Max 1.000 - LowerMedIncome Dummy: Mean 0.400; SD 0.490; Min 0.000; Max 1.000 - UpperMedIncome Dummy: Mean 0.105; SD 0.306; Min 0.000; Max 1.000 ### Instruments (selected instrument summary statistics) - EasternReligionFraction1900: Mean 0.060; SD 0.206; Min 0.000; Max 0.990 - HinduFraction1900: Mean 0.024; SD 0.111; Min 0.000; Max 0.816 - JewishFraction1900: Mean 0.006; SD 0.016; Min 0.000; Max 0.090 - LagEducation: Mean -4.111; SD 2.021; Min -12.18; Max -1.024 - LagInvestment: Mean 2.680; SD 0.554; Min 0.750; Max 4.515 - LagInflation: Mean 14.205; SD 23.796; Min -3.079; Max 270.65 - Landlocked (dummy): Mean 0.137; SD 0.344; Min 0.000; Max 1.000 - lLand (log land area): Mean 12.627; SD 1.577; Min 9.131; Max 16.048 - lPop (log population): Mean 9.671; SD 1.376; Min 6.473; Max 13.978 --- ### Diversification and volatility — literature synthesis and interpretation - Macroeconomic stability has been instrumental to sustained growth and development in developing countries (paragraph 27). - Export diversification can increase resilience to external shocks by reducing dependence on particular products, especially primary products and commodities which tend to be associated with higher risk (paragraph 28). - Empirical findings from the literature cited: - Hadad et al., (2013): Countries with a more diversified basket of export goods are less likely to be impacted negatively by external shocks. - Stanley and Bunnag (2001): For four Central American countries, greater export diversification leads to lower income instability within the 1974 to 1995 period; effects depend on product combinations and volatility/covariance properties. - Agosin (2007): Shows negative correlation between export diversification and export growth variance; suggests diversification could influence growth by reducing variance of export growth. - Koren and Tenreyro (2007): Productive structure moves to less volatile sectors and sectoral concentration declines as countries develop. - Mobarak (2005): Finds higher levels of diversification lower growth volatility. - Bertinelli et al. (2009): Using a panel of developing countries to estimate the trade-off between export earnings and variability based on modern portfolio theory, they find welfare gains from export diversification structure. *International Monetary Fund — Appendix Table A3 and Data Appendix (from the document "DIVERSIFICATION AND TRANSFORMATION IN LICS—BACKGROUND NOTES")* ### 29. With this in mind, the focus is on developing countries and three key questions. ### _030514a - 29. With this in mind, the focus is on developing countries and three key questions. ### Major questions and scope - Focus: developing countries (emerging and developing country group using the IMF WEO classification). - Three key questions: - Are episodes of significant, sustained diversification associated with increased macroeconomic stability? - Does export diversification of products help reduce growth volatility, and is this effect through extensive or intensive diversification? - Does output diversification have an effect on reducing growth volatility? ### Stylized facts on volatility and diversification - Developing countries exhibit a higher level of growth volatility; growth volatility declined after 1995 in both EMs and LICs but remains higher than in advanced countries. - The Theil index is used as a measure of concentration for exported products; a lower Theil index indicates higher diversification in exports. - Observed relationships: - LICs and EMs have higher output volatility and a higher Theil index than AMs. - Positive relationship between higher output volatility and lower diversification, particularly for LICs. - Industry diversification helped attenuate the impact of the global financial crisis (da Costa Neto and Romeu, 2011). - Large diversification spurts: - A total of 61 diversification spurts in 51 developing countries identified in the post–1962 period. - Spurts occurred more frequently in the 1960s and 1990s and were evenly distributed across regions (after controlling for the relative number of countries). - For the sample as a whole, spurts lasted 13 years on average; in the East Asia and Pacific region, spurts lasted 20 years on average. - Diversification spurts are associated with a reduction in the volatility of output growth in developing countries. - In LICs, growth volatility decreases 1.5 percentage points in the wake of diversification spurts. ### Data, variables, and estimation approach - Data period: 1962-2010 for export diversification analyses; 1970-2010 for output diversification analyses (value added data start in 1970; seven sectors available). - Dependent variable: Volatility of GDP per capita growth, calculated as the standard deviation of GDP growth using a five-year window (data are five-year averages; t denotes each five-year period). - Diversification measures: - Total Theil index and Herfindahl index (HFI) capture overall level of export-product concentration (lower values indicate higher diversification). - Extensive margin: exports of new product lines. - Intensive margin: a more balanced mix of existing products (within Theil). - Lower values for all four indices indicate a higher level of diversification. - Controls: trade openness (total exports and imports as a share of GDP), terms-of-trade volatility, inflation volatility, exchange rate volatility; interaction terms between diversification index and trade openness included in some specifications. - Estimation: two-step System GMM (Arellano and Bond, 1991) to address dynamic specification and lagged dependent variable bias; period dummies and a constant included but not reported. ### Empirical results — export diversification and volatility (developing countries and all countries) - General finding: export diversification helps reduce growth volatility. - Interpretation of signs: a lower Theil index (total, intensive, or extensive) or lower HFI means higher diversification; positive coefficients on diversification measures imply higher diversification is associated with lower volatility. - Table 3 (Developing Countries, 1962-2010, System GMM) highlights: - Lagged volatility coefficients reported across columns (examples): 0.459***, 0.629***, 0.445***, 0.222*, 0.0299, 0.330**, 0.113, 0.217, 0.122, 0.0379, 0.0191, -0.00322. - Theil index coefficients: 0.823**, 0.834**, 0.969*, -0.865. - HFI index coefficients: 5.605**, 5.022**, 8.178**, 1.324. - Within Theil (intensive) coefficients: 1.137*, 0.961*, 1.549**, -4.048*. - Trade openness coefficients include: 0.00567, 0.000810, -0.0126, -0.0477, 0.0147, -0.150**. - Interaction terms: Within Theil x Openness = 0.0515**. - Terms of trade volatility example coefficients: -0.0960***, -0.0544, -0.0684*, -0.0796**, -0.0738*, -0.0207. - Exchange rate volatility example coefficients: 0.0462, -0.00562, 0.0120, 0.0454**, 0.0470, 0.0410. - Observations by column examples: 886, 925, 886, 864, 900, 864, 289, 292, 289, 288, 291, 288. - Number of id examples: 133, 141, 133, 132, 140, 132, 55, 56, 55, 55, 56, 55. - Hansen test p-values examples: 0.139, 0.188, 0.0763, 0.108, 0.162, 0.0444, 0.652, 0.621, 0.522, 0.629, 0.471, 0.194. - A-B AR(1) test p-values examples: 0.000635, 0.0338, 0.000605, 0.00440, 0.107, 0.00393, 0.0254, 0.0201, 0.00621, 0.0255, 0.0608, 0.0308. - A-B AR(2) test p-values examples: 0.695, 0.574, 0.775, 0.441, 0.454, 0.672, 0.931, 0.978, 0.856, 0.922, 0.892, 0.785. - Table 4 (All Countries, 1962-2010, System GMM) highlights: - Lagged volatility coefficients examples: 0.434***, 0.629***, 0.445***, 0.245*, 0.0695, 0.277*, 0.465***, 0.562**, 0.562**, 0.220*, 0.152, 0.259**. - Theil index coefficients examples: 1.845***, 0.906, 0.714, -1.397*. - HFI index coefficients examples: 3.196*, 4.814*, 7.487**, -12.35. - Within Theil coefficients examples: 1.882***, 0.930*, 0.810, -2.262. - Trade openness coefficients examples: 0.0131, 0.0149, 0.00698, -0.0399*, 0.00172, -0.0499. - Interaction terms: Theil x Openness = 0.0237**; HFI x Openness = 0.193*; Within Theil x Openness = 0.0289*. - Observations examples: 1,116; 1,350; 1,116; 1,094; 1,144; 1,094; 448; 536; 451; 450; 538; 450. - Number of id examples: 163, 179, 163, 162, 173, 162, 81, 85, 81, 81, 85, 81. - Hansen test p-values examples: 0.176, 0.124, 0.0998, 0.127, 0.0950, 0.0298, 0.314, 0.470, 0.638, 0.182, 0.419, 0.318. - A-B AR(1) test p-values examples: 0.000571, 0.000303, 0.000304, 0.00372, 0.0363, 0.00310, 0.0169, 0.00694, 0.00827, 0.0301, 0.0456, 0.0185. - A-B AR(2) test p-values examples: 0.710, 0.979, 0.802, 0.432, 0.458, 0.539, 0.929, 0.900, 0.642, 0.688, 0.850, 0.626. ### Interpretation and robustness - Effects persist after including trade openness and other control variables: - In developing-country regressions (Table 3 columns (4)–(6)), HFI and intensive Theil remain significant with positive signs; total Theil positive but not significant. - With additional controls (terms-of-trade volatility, exchange rate volatility, inflation volatility), only HFI remains positive and significant. - Results for all countries are similar. - Interaction of diversification and trade openness: - Evidence that export diversification reduces growth volatility more when a country is more open to trade. - Interaction results mirror Haddad et al. (2013): negative coefficients on trade openness and concentration measures and a positive, significant coefficient on the interaction term of trade openness and diversification. - For developing countries: signs align with Haddad et al. (2013) but coefficients for trade openness, total Theil, and HFI are not significant. - For all countries: trade openness and HFI coefficients not significant; total Theil index significant; intensive Theil not significant. - Interaction coefficients reported: e.g., Theil x Openness = 0.0237** (all countries); Within Theil x Openness = 0.0515** (developing countries). - Graphical representation: positive relationship between trade openness × diversification interaction and output growth volatility illustrated in Figure 7. ### Output (real) diversification results - Output diversification measures based on UN value added data (seven sectors; data start in 1970). - Table 5 (1970-2010, System GMM for output diversification) key reported values: - Lagged volatility examples across columns: 0.0703, 0.0939, 0.183, 0.141, 0.211, 0.247*, 0.191, 0.130, -0.141, -0.264. - Theil index, Real Sectors examples: 13.03, 9.323*, 4.045, -13.53**, 23.23**. - HFI index, Real Sectors examples: 26.76, 19.53*, 1.419, -35.00**, 59.03**. - Trade openness examples: 0.00204, -0.00696, -0.00426, 0.000764, -0.0198, -0.0683***, 0.0190, 0.0443. - Interaction terms: Real Theil x Openness = 0.138***; Real HFI x Openness = 0.337***. - Exchange rate volatility examples: 0.0202, 0.0256, 0.0520**, 0.0584*. - Observations examples: 803, 803, 294, 294, 1,004, 1,004, 459, 459, 176, 176. - Number of id examples: 143, 143, 55, 55, 174, 174, 81, 81, 13, 131. - Hansen test p-values examples: 0.468, 0.122, 0.490, 0.441, 0.0549, 0.0120, 0.863, 0.938, 0.478, 0.492. - Key result: interaction term of real diversification and trade openness is significant; a country that is more diversified in output and has higher trade openness has lower growth volatility. ### Conclusions and policy-relevant findings - Export diversification matters for macro-stability in developing countries, particularly for vulnerable LICs where increasing export diversification helps reduce growth volatility. - Diversification can occur through extensive or intensive margins or both; empirical results emphasize the importance of intensive diversification for reducing volatility. - Increased intensive diversification (equalizing export shares of the current export basket) is especially important for lowering output growth volatility. - Mechanism: when an economy becomes less concentrated in specific products—especially primary commodities with volatile prices or high demand volatility—the country can experience decreased growth volatility. - There is also evidence that output diversification (real-sector diversification) reduces growth volatility when combined with greater trade openness. *Prepared by Ke Wang; sections prepared by Ho Giang as noted in the source.* ### 43. Notwithstanding significant heterogeneity across LICs, some key patterns emerge. ### _030514a - 43. Notwithstanding significant heterogeneity across LICs, some key patterns emerge. ### Key patterns and mechanisms - There is a tight relationship between sectoral productivity growth and structural transformation/diversification. - Two broad equilibria are identified: - Virtuous cycle: rapid sectoral productivity growth, productivity-enhancing labor reallocation, and increasing export diversification. - Less optimal equilibrium: slow growth, stagnant productivity, little diversification, and productivity-reducing sectoral shifts. - Structural transformation is not automatic and is strongly influenced by the policy, business, and institutional environment. - Structural reform measures shown to be effective in low-income countries include removing trade barriers and reforming the banking and networks industries, which can kick-start sectoral productivity growth and structural transformation. ### Evolution of sectoral shares and employment - Over the past four decades in LICs: - The agricultural share in total value added has continued to decline. - The service share has increased commensurately. - The share of manufacturing has not changed dramatically and remains at a low level (about 12 percent) for the average LIC. - Regional and sectoral patterns: - The agricultural employment share declined between 1990 and 2007 across regions, but the decline was more pronounced in Asia and less so in SSA. - In the average SSA country, over 60 percent of the workforce continues to be employed in agriculture despite the sector’s low value-added share, indicating very low agricultural productivity. - The service sector expansion in many LICs is characterized by low-skilled, less productive activities (retail, social, personal services), although construction, transportation, and communication have recently been gaining value added shares from a low level. - Manufacturing has gradually moved toward higher skill-intensive activities, but the majority of manufacturing value added continues to be generated by low-skilled industries such as food, clothing, and footwear. ### Sectoral productivity dynamics - Sectoral productivity growth in LICs experienced a strong rebound in the 2000s, with the productivity surge most pronounced in agriculture and services. - Drivers of the 2000s productivity surge included marked improvements in the terms of trade, robust investment rates, and significant reforms in real and financial sectors. - Countries that experienced greater reductions in agricultural shares on average exhibited strong and broad-based productivity growth at the sector level, in marked contrast to those that have not. - Regional disparities in sectoral productivity: - Recent productivity performance was strongest for countries in Asia and ECA (Europe and Central Asia). - Performance was mixed in LAC, MENA, and SSA. - LAC economies registered high productivity growth in agriculture but productivity declined in industry and services. - For MENA and SSA, agriculture and services performed well, while industry productivity stagnated. - Productivity gaps with advanced economies: - Agriculture in SSA exhibits the widest productivity gap, averaging less than five percent of U.S. level (labor productivity measured in constant 2000 PPP US dollars), with the gap remaining roughly unchanged over the fifteen-year period reported. - Relative labor productivity in construction in SSA is about 25 percent of U.S. level. - Manufacturing productivity gaps have widened relative to the U.S. in many SSA economies. - Mining productivity often exceeds that in the U.S. in SSA due to the capital-intensive, resource-driven nature of the sector. ### Manufacturing composition and employment - In 2006, manufacturing in LICs was concentrated in low-technology, labor-intensive activities: - Food and beverages accounted for 32.0 percent of total manufacturing value added and 21.9 percent of manufacturing employment. - Other notable contributors to value added included mineral products (10.2 percent), textiles (7.4 percent), chemical products (7.2 percent), and basic metals (6.6 percent). - Employment shares differ from value added shares; for example, food & beverages contributed 32 percent to value added but 22 percent to employment (figure labels). - Average productivity is relatively higher in mineral products, chemical products, and basic metals, where value added contribution exceeded employment contribution. - Reliable information on informal manufacturing activities is limited; the informal manufacturing sector may be substantially larger than the formal manufacturing sector analyzed. ### Aggregate productivity gains and decomposition - LICs that experienced larger reductions in agricultural share and greater export diversification reaped economy-wide productivity gains. - Larger reductions in the agricultural share during 1990-2010 are associated with faster aggregate labor productivity increases. - Economies with more diversified export bases registered stronger productivity gains. - Decomposition of aggregate labor productivity growth (1990-2007) shows heterogeneity across regions: - Inter-sectoral labor reallocation (sectoral shifts) was most productivity-enhancing in Asia, but contributed negatively to aggregate productivity growth in LAC and SSA. - Within-sector productivity growth was the main driver of economy-wide productivity growth overall. - Sectoral drivers by region: - Asia’s LICs: tradable sectors such as agriculture and manufacturing were primary sources of productivity growth. - LAC, ECA, and to an extent SSA: mining played a relatively dominant role. - SSA: services, particularly wholesale and retail trade, were important contributors to aggregate productivity growth during 1990-2007. - Productivity-enhancing labor reallocation is closely linked to reduced employment in agriculture: - Countries where sectoral shifts contributed positively to aggregate productivity growth (examples: Vietnam, Ethiopia, Albania) experienced labor migration out of agriculture into construction, manufacturing, and services. - Countries with productivity-reducing sectoral shifts (examples: Nigeria, Zambia, Bolivia) saw increases in agricultural employment share or shifts toward less productive services at the expense of more productive industries. ### Correlates of productive sectoral shifts - High agricultural productivity growth and a diversified export base are associated with productive sectoral shifts: - Countries benefiting from structural transformation tend to have high productivity growth in agriculture during the period studied. - The contribution of sectoral shifts to aggregate productivity tends to be higher in more diversified economies (measured by the Theil index) and in non-commodity exporters compared to commodity exporters. - The analysis acknowledges possible reverse causality: labor moving out of agriculture would, ceteris paribus, increase productivity of remaining agricultural workers. *DIVERSIFICATION AND TRANSFORMATION IN LICS—BACKGROUND NOTES, INTERNATIONAL MONETARY FUND.* ### 55. Productivity gaps between agriculture and non-agricultural sectors remain large in ### _030514a - 55. Productivity gaps between agriculture and non-agricultural sectors remain large in ### Productivity gaps: levels and patterns - Agricultural productivity as a percentage of labor productivity in non-agricultural sectors: - Averaged about 30 percent in 2005 for LICs—little changed from a decade ago. - Compared to 40 percent in emerging market economies. - Cross-country patterns and correlates: - Less diversified economies (by exports) tend to exhibit much wider productivity gaps than more diversified economies. - Commodity exporters tend to have larger productivity gaps on average. - There is a relationship with development level: LICs tend to have lower relative agricultural productivity. - McMillan and Rodrik (2011) pattern replicated: a U-shaped relationship across countries and over time, so relative agricultural productivity may initially worsen as non-agricultural sectors expand before improving at higher development levels. ### Drivers of sectoral productivity and structural transformation — model and data - Objective: Identify policy and institutional variables associated with within-sector productivity growth and sectoral shifts in LICs. - Dependent variables: - Log of average labor productivity in agriculture, manufacturing, or services (annual sectoral productivity growth rate). - Alternatively, (real) value added or employment shares of agriculture, manufacturing, or services (for sectoral shifts). - Key explanatory variables considered (lagged one period, entered one-by-one in regressions): reform indices (international trade, domestic financial sector, product/network markets, FDI), labor cost (average tax wedge), labor market and business regulation (Fraser regulation index), infrastructure stock (log per capita road network), and education (tertiary education measured as percentage of high school). - Convergence control: initial sectoral productivity/share gap with the US included. - Estimation: - Two-step GMM estimator (Arellano and Bond, 1991), using first differences and lags of the dependent variable as instruments. - Panel: 28 LICs for the period 1995-2010. - De-trended measure of the dependent variable with country-specific linear trends. - Caveats: analysis focuses on short-term impacts of de jure reforms and estimates sectoral productivity growth and sectoral shifts separately, abstracting from dynamic feedbacks between them. ### Results — sectoral productivity (high-level findings) - Different sectors respond to different reforms; key associations identified: - Agriculture: - Removing tariff barriers to international trade is associated with higher agricultural productivity. - Domestic financial sector reform in the area of interest rate controls has a positive effect on agricultural productivity. - The effect of other reform measures, including agricultural reform, on agricultural productivity is not statistically significant in the estimates. - Manufacturing: - Domestic financial sector reforms, capital account liberalization (FDI), improvement in road infrastructure, and tertiary education are associated with higher manufacturing productivity. - Rationale: financial reform improves capital allocation; services reform can have spillovers into manufacturing; road connectivity reduces trade frictions; tertiary education supports skill-intensive manufacturing. - Services: - Liberalization of network industries, specifically telecommunication, is associated with productivity gains in the service sector. - Road infrastructure improvement also has a positive effect on service productivity growth, notably for distribution services. - Selected quantitative estimates from Table 6 (coefficients with reported standard errors and significance): - Trade liberalization (tariff) effect on agricultural productivity: 0.175 [0.093]*. - Trade liberalization (tariff) effect on manufacturing productivity: 0.053 [0.096]. - Trade liberalization (tariff) effect on services productivity: -0.177 [0.107]*. - Removal of interest rate control effect on agriculture: 0.640 [0.275]**. - Liberalization of FDI effect on manufacturing: 0.369 [0.214]*. - Networks reform effect on services: 0.138 [0.063]**. - Road infrastructure effect on manufacturing: 0.091 [0.028]***. - Tertiary education effect on manufacturing: 0.017 [0.009]*. - Tertiary education effect on agriculture: -0.026 [0.014]*. - Note on variables and notation in Table 6: - Reform indices normalized between 0 and 1. - Fraser regulation index ranges from 0 to 10 (higher = less regulation). - Road infrastructure is the log of per capita road network. - Tertiary education is the percentage of high school. - Estimation: two-step GMM with 3 lags of dependent variable as instrument. All specifications include year fixed effects. Significance: * 10%, ** 5%, *** 1%. ### Results — sectoral shifts (high-level findings) - Policies and institutions also have direct associations with sectoral shifts (in both real value added and employment shares): - Trade liberalization (removing tariff barriers) is associated with resources moving from agriculture to manufacturing and services; structural shifts occur in both employment and value added (manufacturing especially). - Infrastructure and regulatory environment matter for labor mobility and skills supply: - Liberalization of the electricity market, easing credit, and labor market and business regulations (Fraser regulation index) are linked to higher industry employment share. - Improving tertiary education increases the value added shares of manufacturing and services. - Theoretical context: trade openness can change relative prices and induce resource reallocation to sectors with comparative advantage; empirical cross-country evidence is mixed and depends on episode selection and sectoral aggregation, but several case studies corroborate the role of trade in moving labor into manufacturing. ### Conclusions and policy implications - Summary of stylized facts: - Sectoral shifts and sectoral productivity growth are central to growth and development. - Structural transformation in LICs remains low and uneven; some LICs (notably in Asia) have experienced robust productivity growth and labor movement out of agriculture, while others remain concentrated in low value-added activities. - Structural transformation does not occur automatically—countries need conducive structural and institutional settings to enable productivity growth and labor mobility. - Policy measures found to provide short-term impulses to sectoral productivity and sectoral shifts in LICs over the 1995-2010 sample include: - Removing tariff barriers to international trade. - Reforming the domestic financial sector (including removing interest rate controls). - Liberalizing networks (notably telecommunications) and other network industries. - Liberalization of FDI/capital account (in some cases). - Improving road infrastructure. - Enhancing tertiary education. - Easing regulatory burdens and improving electricity market liberalization and credit access. - Sector-specific focus is important: different sectors require different policy priorities to achieve productivity gains and to facilitate structural transformation. *Source: IMF staff calculations and analysis in “Diversification and Transformation in LICs—Background Notes,” section on productivity gaps, drivers, model, results, and conclusions.* ### 69. Economic development is associated with the transformation of a country’s economic ### _030514a - 69. Economic development is associated with the transformation of a country’s economic ### Definitions and conceptual framework - Economic transformation has two dimensions: - Horizontal: diversification into new higher value added sectors. - Vertical: quality upgrading—producing higher quality (and generally higher priced) products within existing sectors. - Producing higher-quality varieties of existing products builds on existing comparative advantages to boost export revenues and productivity. - Both dimensions are important for LICs; some LICs require horizontal diversification to access large quality-upgrading gains, while others can harness quality upgrading within current sectors. ### Measurement and data - Quality is not directly observed; unit values (average export prices) are a noisy proxy because they reflect production costs, pricing strategies, and destination composition. - A multi-level export quality database (Henn, Papageorgiou, and Spatafora, 2013) corrects unit values for these factors and is available as an online toolkit to Fund economists. - Database coverage and structure: - 178 countries and 851 products over 1962–2010. - At the SITC 4-digit level: more than 20 million product-exporter-importer-year observations. - Quality estimates also at SITC 3-, 2-, and 1-digit levels; country-level BEC classification; and three broad BEC sectors (agriculture, manufactures, non-agricultural commodities). - All quality estimates normalized to the world frontier quality, assumed to be the 90th percentile in each product-year combination. - Resulting quality values typically range between 0 and 1.2. - Normalization is repeated at each aggregation step; rising country quality implies upgrading faster than the world on average. ### Stylized facts on quality upgrading (aggregate and regional) - Regional drivers and patterns: - Quality upgrading among low- and lower-middle-income countries was most strongly driven by East Asian success in the manufacturing sector. - Apparel served as an important first beachhead in manufacturing, with notable quality upgrading in East Asia. - Sub-Saharan Africa shows some gradual manufacturing quality upgrading since 1990. - Latin America’s manufacturing quality has been stagnant; agricultural quality fell until the early 2000s. - Agriculture shows indications of quality upgrading since 2000 in East Asia and Middle East and North Africa. - Fragile vs non-fragile LICs: - Only non-fragile LICs have upgraded export quality considerably. - In the late 1980s, average export quality of non-fragile LICs was among the lowest five percent worldwide; since then their average quality rose to three-fourths of the world frontier quality, with manufacturing particularly important. - Fragile LICs have not converged in quality and remain among the lowest quality exporters. - The fragile/non-fragile pattern may partly underlie differences in annual GDP per capita growth since 1990: 3.4 vs. 1.1 percent. - Quality tends to evolve gradually. - Quality and prices: - Quality increases lead to higher export prices at the quality levels at which most LICs currently produce. - Data suggest quality increases translate into export price increases until a country’s quality level reaches about 80-85 percent of the world frontier. - For most LIC products (below that 80-85 percent level), quality increases would likely result in terms of trade improvements. - Quality increases beyond 80-85 percent tend not to drive prices higher, possibly because higher production efficiency keeps prices stable. - Quality–price correlation is particularly strong in agricultural goods. ### Quality upgrading and economic growth - Quality upgrading is associated with growth: - Quality upgrading is particularly rapid during early stages of development; quality convergence is largely completed as a country reaches upper-middle-income status. - LICs that converged fast in quality during 1995-2010 typically grew one percentage point per year more than countries that converged slowly. - The link between quality upgrading and growth is strongest for manufacturing. - Quality upgrading in manufacturing is accompanied by an increasing share of the sector in LIC economies. - Agriculture: - Ample quality upgrading opportunities also exist in agriculture, typically associated with rebalancing of the sector toward higher value products. - Increasing productivity in agriculture typically leads the sector to set free resources for other sectors. ### Policy-relevant determinants and implications - Once a country has entered an export product, quality tends to converge unconditionally over time—implying policy focus should be on facilitating entry into new markets and products. - Policy priorities: - Facilitate entry into new markets for domestic and foreign firms, especially where products have large quality-upgrading potential. - Improve institutional quality and human capital to accelerate quality upgrading. - There is no evidence that lack of demand for quality in LICs' and lower-middle-income countries' existing destination markets constrains quality upgrading—opening new export markets could take a lower initial priority. - For fragile LICs, horizontal diversification combined with quality upgrading offers the most promise given current specialization patterns. ### Sector-specific findings for LIC export sectors - Fruits and vegetables: - Many African LICs (driven by non-fragile LICs) experienced new export success in fruits and vegetables during the last decade. - Non-fragile LICs strongly upgraded quality in this sector since 1990, yielding steep world market share increases starting about five years after quality rises; later boosted by price reductions. - Success often driven by integration into regional or global value chains through capacity building (example: Zambia becoming a net exporter following entry of foreign-owned supermarket chains). - Fragile LICs experienced declining quality and failed to increase market share in the same period. - Coffee, tea, cocoa, and spices: - Non-fragile LICs have turned quality increases into market share gains since early 2000s. - Fragile LICs have been reversing previous quality declines but have not yet achieved market share gains. - Apparel: - Apparel is typically one of the first manufacturing sectors a country enters and has been particularly important for LICs. - Non-fragile LICs (notably Asian LICs such as Bangladesh and Vietnam) recorded quality increases since the late 1980s and drastic market share increases; East Africa has also increased market share from a low base. - Fragile LICs have been upgrading apparel export quality since 2000, but market share has not yet responded. - Historical dynamics: some countries (e.g., Korea and Thailand) entered and subsequently withdrew from apparel as they developed further; China’s success involved substantial quality increases against stable prices. ### Quality ladders and potential for upgrading - Sectoral quality ladders (SITC one-digit level) vary considerably in length; a country’s relative position also varies across sectors. - Non-fragile LICs: - Have diversified into manufactures, with apparel (miscellaneous manufactured articles) especially important in Asian LICs. - Typically occupy relatively low rungs on quality ladders—implying considerable quality upgrading potential remains. - Henn et al. (2013) show large potential for quality upgrading even in Asian LICs, including apparel. - Fragile LICs: - Remain heavily specialized in minerals, crude materials, and food/live animals. - West Africa and Central Africa remain highly dependent on these sectors. - Within existing sectors, fragile LICs occupy generally lower rungs on quality ladders than non-fragile counterparts—indicating potential for quality upgrading in agricultural and commodity sectors, albeit subject to climatic and geological limitations. *Source: IMF—The Role of Structural Transformation and Diversification—Background Notes (excerpts)._030514a* ### 85. Horizontal diversification may also become paramount again at later development ### _030514a - 85. Horizontal diversification may also become paramount again at later development ### Horizontal diversification and late-stage development - Horizontal diversification may become a precondition for further quality upgrading in more mature exporters. - Example: Malaysia is heavily specialized in electronics exports (a subcategory of the machinery and transport equipment sector) and is already approaching the world frontier quality in this sector; further quality increases in this sector tend to be offset by price declines for constant-quality products. - Horizontal diversification can occur: - across SITC1 sectors, and - across other subsectors within machinery and transport equipment. - Countries positioned between Asian LICs and Malaysia (e.g., China) have realized some quality upgrading but retain additional potential. - Increasing value added in existing exports can come from engaging in more sophisticated tasks beyond assembly (offshoring literature examples: Grossman and Rossi-Hansberg, 2008; Baldwin and Robert-Nicoud, 2010). ### Export quality and income levels - Export quality and income per capita are correlated, particularly strongly in early stages of development. - Quality upgrading is particularly rapid until GDP per capita reaches $10,000. - Quality convergence then continues at a diminishing rate and is largely complete by the time GDP per capita reaches $20,000. - Among high-income countries, average export quality levels vary only within a narrow band. - Pattern holds for aggregate exports and separately for manufacturing, agriculture, and non-agricultural commodities. - The quality measure used evaluates the quality of exported goods but not domestic input content; it can be misleading when low-value assembly uses high-quality imported intermediates to generate high-quality exports. ### Quality upgrading across sectors and countries - There is wide variation in export quality levels, especially among low-income and middle-income exporters. - Countries with low average quality have scope to upgrade within existing sectors. - Countries with relatively high export quality but low incomes may be in sectors with short quality ladders or low productivity and could benefit from horizontal diversification. - Agriculture shows scope for quality upgrading as countries develop—quality of agricultural products and commodities increases substantially, often reflecting shifts toward more processed products. - Small states: - Follow similar patterns: quality rises with income sharply below $10,000. - Have upgraded manufacturing quality considerably since the early 2000s, approaching world frontier quality in manufactures. - Agricultural export quality in small states remains low, though fish and fish preparations subsector saw respectable upgrading. ### Quality upgrading and growth (1995-2010 and 2000-10 findings) - Quality upgrading is associated with higher GDP per capita growth across agriculture, manufacturing, and commodities (1995-2010), with sectoral differences: - Agriculture: a 0.1 quality increase during 1995-2010 was associated with additional annual GDP per capita growth of 0.5 percentage points. - This result holds for LICs, with growth benefits similar between fragile and non-fragile LIC subgroups. - No evident growth benefit in LICs from quality upgrading in the commodity sector. - Manufacturing: a 0.1 quality increase during the 15 year period was associated with additional annual growth of one percentage point. - The strength of the effect is similar for LICs. - Manufacturing quality upgrading is accompanied by increasing importance of the sector in LIC economies (2000-10): - Quality upgrading in manufacturing correlates with growth of manufacturing output (IMF supply-side disaggregated GDP for African and Asian LICs, 2000-10). - Quality upgrading in agriculture is associated with lower growth of the agricultural sector, implying redeployment toward manufacturing and services. - Heterogeneity in quality upgrading: - A subset of low- and lower-middle-income countries experienced fast quality convergence, mostly since the late 1990s. - In Asia: China and Indonesia had rapid upgrading starting mid-1980s; India, Vietnam, Cambodia started rapid upgrading in the late 1990s. - In Africa: fast convergers include South Africa, Ghana, Liberia, Sierra Leone. - Other countries experienced stagnant or slower convergence (e.g., Pakistan, Philippines, Morocco, Cote d'Ivoire), with some showing tentative upgrading only from the mid-2000s (e.g., Bangladesh, Sri Lanka, Cameroon, Kenya). - Growth differentials: - Annual GDP per capita growth was more than one percentage point higher on average in faster than in slower quality LIC convergers during 1995-2010. - The additional growth dividend is slightly larger when lower-middle-income countries are included. - For Africa only, the growth dividend from quality upgrading was lower (though still in excess of one-half a percentage point per annum). - Small states that were fast convergers realized an additional annual per capita growth of two percentage points relative to slow converging peers. - Fast quality convergers during 1995-2010 had experienced lower growth before 1996, indicating a structural break in the mid-1990s associated with different quality upgrading experiences. ### Determinants of quality upgrading - Product-level convergence: - Quality of individual products tends to converge across countries unconditionally; speed of convergence is relatively fast at 4-5 percent per annum. - Within manufacturing, convergence is faster at about 7 percent per annum. - Despite product-level convergence, country-level overall export-basket convergence can be limited by horizontal diversification into new products entered at low quality. - Policy-relevant determinants associated with faster product quality growth include: - Improved institutional quality. - Higher human capital. - Liberal trade regimes. - Liberal agricultural policy. - Liberalized domestic financial markets. - Policy implication from unconditional convergence: facilitating entry into new markets (for domestic and foreign firms) supports automatic quality convergence and can yield substantial quality increases when combined with institutional and human capital improvements. *Source: DIVERSIFICATION AND TRANSFORMATION IN LICS—BACKGROUND NOTES (selected sections E–G, export quality and determinants).* ### 101. Developing countries’ potential for quality upgrading does not appear to be limited ### 101. Developing countries’ potential for quality upgrading does not appear to be limited ### Demand in destination markets and upgrading potential - Lower-income countries tend to serve markets that import lower-quality products, but "the differences are not substantial enough to act as a constraint on quality upgrading." - On average, the poorer the exporter, the greater the gap between its export quality and the average quality demanded by its trade partners in those products that the exporter sells to them. - Within LIC subgroups: - Fragile LICs show a larger gap than non-fragile peers. - LICs in all regions in Africa show a larger gap than more advanced Asian LICs. - Upgrading potential also remains for small states in their existing destination markets. - Empirical note referenced: Henn et al., (2013) (results not reported here). ### Fast vs. slower quality convergers and destination-market constraints - For countries with slower convergence, export quality remains substantially lower than the average quality of their trade partners’ imports, especially for African countries. - Fast convergers have closed the gap between quality exported and quality demanded somewhat. - For both fast and slower convergers, "quality demanded by trading partners does not yet pose a constraint to upgrading." - Policy implication drawn: focus on creating a domestic environment broadly conducive to quality upgrading; facilitating entry into higher-quality export markets is a less urgent priority. ### Stylized empirical conclusions on quality, income, and upgrading (paragraphs 103–107) - "Quality is strongly correlated with income per capita, and quality upgrading is associated with per capita growth." - Pace of quality upgrading by development stage: - Particularly rapid during early stages of development, until a country reaches a GDP per capita of about $10,000. - Convergence in export quality continues at a slower pace until GDP per capita reaches $20,000. - Quality upgrading levels off thereafter (after $20,000). - "Overall, those countries and time periods that have witnessed faster growth in quality have also experienced faster growth in GDP per capita." - Cross-country and regional differences: - Product quality in sub-Saharan Africa and South Asia is lower, and has been growing more slowly, than in East Asia. - Considerable heterogeneity within regions (e.g., quality rising far more rapidly in Ghana or South Africa than in Cote d’Ivoire or Cameroon). - Fragile LICs, often natural-resource based, need horizontal diversification across sectors before harvesting quality upgrading opportunities, compared with non-fragile peers. - Some middle-income countries that increased quality sharply in the past (e.g., Malaysia, and to a lesser extent China) "may now have less scope left to upgrade quality within existing export sectors." - Sectoral implications: - "Sectors with long 'quality ladders' may hold particular potential given that, within any given product line, quality converges to the 'frontier' over time." - Removing barriers to entry into new sectors could boost growth by increasing potential for future quality upgrading. - Diversification across sectors and quality upgrading are complementary. - Agriculture: - Substantial potential for quality upgrading in agriculture for many LICs, and agricultural productivity-enhancing upgrading typically releases labor toward manufacturing and services. - There are success stories of African countries expanding exports in agricultural products (often in fruit, vegetables, or flowers they did not produce before), but these must become much more widespread to deliver major development impacts. - Key constraints include efficient air transport in the success stories of Kenyan and Ethiopian fresh cut flowers. - Addressing these constraints will likely involve close cooperation with the private sector and openness to foreign direct investment to facilitate capacity building through integration into global value chains. - Institutional quality and human capital: - "Both institutional quality and human capital are strongly associated with the pace of quality upgrading." - Empirical note from Henn et al., (2013): a one standard deviation increase in institutional quality or in human capital is associated with, respectively, a 0.3 and a 0.2 standard deviations increase in the pace of quality upgrading. - Demand constraint finding reiterated: - "There is no evidence that lack of demand for quality in a country’s existing destination markets on average constrains quality upgrading." ### Policy recommendations and implications - Emphasize horizontal, economy-wide policies rather than sector-specific interventions to encourage quality upgrading: - Pursue policies that shape a domestic environment with open markets broadly conducive to quality upgrading. - Create new upgrading opportunities through diversification. - Rationale for horizontal policies: - Horizontal policies are preferable given the unpredictability of export successes (e.g., relatively few ‘big hit’ products account for a large share of many developing countries’ exports, per Easterly et al., (2009)). - Specific policy actions implied by the analysis: - Improve institutional quality and human capital to accelerate the pace of quality upgrading. - Remove barriers to entry into new sectors to expand the set of sectors with long quality ladders. - Support agricultural quality upgrading where much LIC labor remains employed, including logistics (e.g., air transport) and integration into global value chains via private sector cooperation and FDI openness. *Source: DIVERSIFICATION AND TRANSFORMATION IN LICS—BACKGROUND NOTES (excerpt).* ### 116. Diversification of exports has followed a similar pattern as that of output, though ### _030514a - 116. Diversification of exports has followed a similar pattern as that of output, though ### Export diversification: broad patterns and heterogeneity - General trend: greater diversification in terms of export products and a move away from agricultural exports (Figure 31A). - Uneven country experiences: - Angola: steady increase in export concentration driven by oil exports; civil war and Dutch-disease effects undermined other exports. - Bangladesh: period of diversification when garments replaced agricultural exports, followed by a re-concentration of exports. - Malaysia: historically the most diversified economy in the sample, but experienced some re-concentration as electronic products increasingly dominated overall exports in the 1990s, and primary commodities regained prominence. - Angola: oil dominance implies 99 percent of exports consist of oil; rising oil price likely obscures some modest export diversification. - Relationship between income and exports: in the sample countries, higher income is associated with higher exports regardless of diversification — countries with the fastest per capita GDP growth also experienced the fastest growth in exports as a share of GDP (Figure 31B). Bangladesh follows this pattern despite decreasing export diversification. ### Product composition shifts and sectoral shares - Share of agricultural in total exports declined in all countries; share of manufacturing increased in most (Figure 31C). - Exception: Malaysia, where agricultural products (largely rubber and palm oil) continue to contribute significantly to total exports, and exports of manufactured goods declined as commodities gained importance. - Share of nonagricultural primary commodities in total exports remained remarkably stable in most countries, except: - Tanzania: gold has become the single largest export product (driven primarily by rapidly rising gold prices). - Within manufacturing: - Share of relatively unsophisticated manufactures (labor-intensive, resource-based and low-skill and -technology) in overall manufacturing exports declined everywhere except Bangladesh, where textiles remain dominant (Figure 31D). - Unsophisticated manufactures gained share in total exports in some countries with rapid total export growth. - Share of more sophisticated manufactures in total manufacturing exports (and to some extent in total exports) increased in most countries, though it declined in Bangladesh. - Malaysia saw a decline in sophisticated manufactured exports in total exports and no expansion in overall manufactured exports, likely related to increasing value of commodity exports. ### Vulnerability and terms-of-trade - Diversification measures per se do not suggest a correlation with vulnerability to terms-of-trade shocks. - The share of non-agricultural primary commodities (mainly fuel, metals, and gold) in total exports appears strongly related to terms-of-trade movements (Figure 31E). - Malaysia: concentration in electronics and electrical category may not capture movement to higher value-added products; standard deviation of terms-of-trade changes over the past two decades for Malaysia has been more than twice as high as for developing Asia as a whole, and is also higher than Vietnam’s. ### Geographic diversification - Exports generally diversified geographically (Figure 31F): share of the top five trading partners in exports declined in all countries since 1981, except Bangladesh (increasing reliance on a few markets). - Angola: geographic trade concentration has decreased but remains high because oil is destined to a few markets. - Caveat: with the price of oil determined in a global market, diversification of export destinations will do little to mitigate price risk. ### Cross-country comparability and limitations - Analysis provides a relatively rough comparison between countries; cross-country comparability comes at the price of loss of granularity that can obscure microeconomic and country-specific trends (paragraph 117). - Important aspects of diversification pertain to specific sub-sectors or require higher-granularity data not available comparably across countries. - Factors behind evolution of diversification and policy lessons are better gleaned at the individual-country level. ### Country experience — Malaysia (selected case study) - Long-run outcomes: - GDP per capita (PPP) rose from less than $900 forty years ago to over $16,000 in 2011. - Exports of goods and nonfactor services (GNFS) increased from about 40 percent of GDP in the early 1970s to more than 100 percent in the mid-2000s. - Phases of diversification: - Primary commodities and agriculture: - 1960s: tin, crude oil, and rubber accounted for two-thirds of Malaysia’s total exports (Figure 32A). - Palm oil emerged as a major new industry; government policies (tax incentives, export promotion, publicly funded research and training, opening new land for cultivation) promoted palm oil cultivation and processing. - Crude oil rose in prominence in the 1970s; share of crude oil in exports peaked at 30 percent in 1984 when manufacturing began to take off. - Industrialization: - Initial import substitution industrialization (ISI) led to manufacturing share rising from 8.5 percent of GDP in 1961 to 16.5 percent in 1975 (Chee, 1987). - Shift in late 1960s/early 1970s to export-oriented, labor-intensive industrial development relying on free trade zones (FTZs). - Foreign investment rose from 2.25 percent of GDP in 1970 to almost 9 percent in 1992, before declining again. - Manufacturing rose from 12 percent of GDP in 1969 to a peak of 31 percent in 2006. - Manufacturing exports rose from 6 percent of total exports to 74 percent over the same time (Figure 32C). - Trade geography broadened: share of top five export destinations declined from 67 percent in 1981 to 51 percent in 2011 (Figure 32D). - Electronics concentration and volatility: - Electronics industry’s share in manufacturing and total exports rose during the 1990s to a peak of over 72 percent and 58 percent, respectively, in 2000 (Figure 32E). - This re-concentration exposed Malaysia to terms-of-trade–type volatility; only in the last decade has the share of electronics in manufacturing and exports declined again. - Lessons and challenges from Malaysia: - Deliberate diversification and export promotion policies can support structural transformation and growth. - Focused state support can spur new industries but also skew economic structure and create vulnerabilities (example: electronics concentration). - As Malaysia reached middle-income status, dominance of manufacturing is declining; further diversification within manufacturing and development of the service sector will be important. - Raises question whether the model of strong state involvement, sector-specific promotion, and FTZs remains suitable for a structurally changed economy. ### Country snapshots (selected points from the subsection introduction) - Vietnam: - GDP per capita (PPP) rose from less than $700 in 1990 to over $3,300 in 2011. - Exports of goods and services rose from about one-third of GDP to 86 percent of GDP. - FDI inflows rose from an annual average of less than 4 percent of GDP in 1990-92 to 7½ percent of GDP in 2008-11. - Agricultural sector declined from 32 percent of GDP in 1990 to 16 percent in 2011; industry and construction rose from 25 percent to 42 percent over the same period (Figure 33A). - Bangladesh: - Exports largely textiles with decreasing export diversification; nonetheless saw rising per-capita output and exports. - Angola: - Exports dominated by oil; 99 percent of exports consist of oil. - Tanzania: - Gold became the single largest export product; nonagricultural primary commodity share changed markedly due to gold. *Sources: UNCTAD, country authorities, IMF staff calculations.* ### 131. Economic reforms were pursued as a response to challenges that arose to the ### Economic reforms were pursued as a response to challenges that arose to the government’s goal of sustaining high growth. ### Launch and content of Đổi Mới (1986 onward) - Đổi Mới launched in December 1986. - Initial reform focus: - Agriculture: individual land use rights recognized; collective assets privatized; production freed from state-set quotas. - Prices and subsidies: between 1987 and 1990, prices were liberalized, the rationing system largely abolished, and subsidies significantly reduced. - External sector: trade and access to foreign exchange were partially liberalized; the official exchange rate devalued; economy gradually opened further to FDI. - Macroeconomic policy: focus on reducing inflation. - Implementation approach: step-by-step, building on newly-gained experience. - Source cited: Van Arkadie and Mallon, 2003 (as presented in text). ### Drivers of early growth and structural change - Key strengths enabling post-reform expansion: - Large, low-cost and relatively well-educated labor force (education levels much higher than in countries of comparable per-capita income). - Strong work ethic and entrepreneurial spirit. - Sizable diaspora of overseas Vietnamese as a source of knowledge and capital. - Proximity to major economic centers in the region. ### Agricultural expansion and early export diversification - Demographics and agriculture: - Around 80 percent of the population living in rural areas. - Agricultural outcomes: - Production of staples, especially rice, increased rapidly. - Sector diversified into cash crops and marine products; production of marine products grew even faster than rice. - Export shares: - Agricultural exports accounted for almost half of all exports in 1995 before their share declined as oil and industrial products became more prominent. - Within agricultural exports, rice’s dominant position gradually eroded as coffee, marine and forestry products gained importance. ### FDI, industrialization, and export transformation - FDI patterns and effects: - Initial FDI concentration in the oil sector; subsequently real estate (including hotels), food processing, and from the mid-1990s both heavy and light industries gained importance as ownership restrictions were lifted. - FDI instrumental in raising exports; share of foreign-invested enterprises in exports grew. - Evolution of export composition: - Exports of oil and heavy industrial products rose first, followed by increases in light industries (textiles, footwear, electronics). - Geographic diversification of exports: - Late 1980s: Soviet Union was most important trade partner. - Later replaced by Japan, other Asian countries, European countries other than Russia, and the United States. - FDI from Asian countries (Korea, Taiwan POC, Singapore and Malaysia) accounted for almost 45 percent of cumulative FDI (up to 2010), while share of exports to these countries remained much smaller. - The United States became a major export destination despite relatively small U.S. direct investment in Vietnam. ### Further reforms after the 1997/98 Asian crisis - Rationale: initial reforms increased productivity and growth but effects faded and growth declined; the 1997/98 crisis exposed reform limits and spurred a second round. - Reform priorities: - Reduce and simplify administrative procedures. - Improve legal standing of private and foreign-invested enterprises; ensure equal treatment of domestic and foreign-invested enterprises. - Further open the economy to trade and competition; reduce cost of doing business. - Reforms in telecommunications and power to reduce communication and electricity costs. - Investment in infrastructure to lower transportation costs. - WTO accession: - Accession to the World Trade Organization (WTO) in January 2007 served as a commitment device to reinforce credibility of reforms. - State role and SOEs: - State granted preferential treatment to some SOEs to build strategic industries (e.g., shipbuilding), though this has not played a major role in diversification. - Note in text: SOEs’ performance improved compared to pre-reform era; however, they tend to lag behind private sector in capital and labor productivity, and increased autonomy has become detrimental to accountability. ### Financial and remittance developments - Post-reform financial changes: - Overall investment and FDI rose rapidly, with a large share of investments going to wholly foreign-owned entities. - Role of joint stock private banks grew, as did credit to the private sector. - Remittances: - Remittance inflows continued to rise; government reforms to encourage remittances included abolishing a 5 percent tax on remittance inflows, discontinuing arbitrary treatment and potential seizure, allowing foreign transfer companies, and improving investment opportunities. ### Lessons and pending challenges for Vietnam - Successive “waves” of supportive reforms sustained diversification and structural transformation. - Diversification was largely a byproduct of broader economic reforms rather than an explicit, singular policy goal. - Current constraints (as of text): - Reform efforts have run their course; credit- and investment-driven growth model reached limits; macroeconomic stability eroded and growth slowed amid the global crisis. - Government response and reform focus moving forward: - Re-emphasize macroeconomic stability. - Prepare new reforms aimed at streamlining public investment and improving SOE performance to boost medium-term growth. - Two key areas for further reform to achieve middle-income status and higher-value-added production: - Invest in higher education to equip workforce for quality and innovation and enable entrepreneurs to enter new production areas. - Rely more on market mechanisms and less on direct administrative controls to steer the economy, while maintaining regulatory oversight to contain negative externalities and systemic risks (e.g., in the financial sector). --- ### Tanzania: reform-driven transformation (selected findings) - Reform chronology and scope: - Comprehensive reforms began in mid-1980s, accelerated from mid-1990s, transforming a state-dominated welfare state to a market economy. - Reforms covered exchange, price and trade liberalization, tax and public finance reforms, modern monetary policy, financial sector liberalization, public sector reforms and privatization of SOEs, and regional integration. - Growth and income outcomes (post-2000): - Growth accelerated from an average of 3 percent during the 1990s to an average of 6¾ percent during 2000-2011. - Per-capita income (PPP basis) more than doubling to $1,610 since 2000. - Labor and total factor productivity increased by more 30 percent in the last decade (as stated in text). - Inflation fell from double to single digits. - Public debt burden declined substantially (partly due to increased foreign aid and debt relief under HIPC initiative and MDRI in the mid-2000s). - Financial deepening and public finance reform: - Private sector credit grew from about 4 percent of GDP in the late 1990s to close to 19 percent in 2011. - Tax policy reforms included introduction of VAT in 1998 and improved tax collections. - Sectoral and export diversification: - Tanzania moved from being the least diversified economy in the sample in 2000 to the most diversified by 2011. - Agricultural share of total output declined from 32 percent in 1998 to 23 percent in 2011. - Agricultural diversification: high-volume cash crops (cotton, coffee) continued to grow; smaller-volume crops (cashew nuts, tobacco, tea) grew faster; production was reoriented toward domestic market needs. - Manufacturing, construction and services (excluding public sector) increased their shares of the economy. - Mining (mainly gold) expanded from 1.6 percent of output in 2000 to 2.7 percent in 2007, then declined to 2.3 percent by 2011. - Gold’s impact on trade was amplified by a rise in gold prices of over 450 percent during the same period. - Reform sequencing and reversals: - Initial reforms included introduction of a crawling peg in 1986 and full exchange rate unification in 1993; domestic price controls lifted by 1991 on almost all products (from more than 400 regulated prices in 1990), except for petroleum and public utilities. - Privatization program launched in 1993; by 2003 most manufacturing and commercial parastatals had been restructured, liquidated or transferred to private ownership. - Reforms were temporarily derailed in the early 1990s but resumed stronger from 1996 onward. - External support: - Increased donor support following reforms helped raise public investment and poverty alleviation. *Source: IMF staff background notes as presented in the supplied content.* ### 150. Gold has become the single largest export, accounting for 9¼ percent of GDP in 2011, ### _030514a - 150. Gold has become the single largest export, accounting for 9¼ percent of GDP in 2011, ### Tanzania — export structure and recent evolution - Gold accounted for 9¼ percent of GDP in 2011. - Gold accounted for 30 percent of total exports of goods and nonfactor services in 2011. - Gold accounted for 43 percent of exports of goods in 2011. - Tanzania became the fourth largest gold exporter in Africa. - Between 2000 and 2011, the volume of gold exports rose by about 250 percent. - Much of the increase in gold exports reflected improved terms of trade, but volume growth was also substantial. ### Tanzania — broader export and sectoral diversification - Overall exports of goods and nonfactor services grew from 14½ percent of GDP in 1997 to 31 percent in 2011. - Manufacturing exports grew from 3 percent of non-gold GNFS exports to 16 percent of non-gold GNFS exports. - Manufacturing exports rose by 5½ percentage points of GDP. - Traditional exports declined from 19 percent to 13 percent of non-gold GNFS exports. - Sectoral change: agricultural share shown as (RHS) in Figure 34A; manufacturing, wholesale trade, construction, and agriculture are tracked in sectoral shares 2000–11 (figures referenced). ### Tanzania — changing geography of trade - The share of the EU as a main trade partner declined between 1991 and 2011. - New partners emerged, notably Switzerland and China, reflecting gold exports. - Regional trade increased, especially with the EAC and South Africa. - Destination-of-exports composition tracked for 1991–2011 (percent of total). ### Lessons and policy implications for Tanzania - Broad economic liberalization over the past 1½ decades yielded significant benefits in growth and diversification. - External factors—rising commodity prices—supported investment in new industries, diversification, and additional government revenue that enabled higher public investment benefiting other industries. - Tanzania still has relatively low income per capita and must: - Sustain recent growth rates. - Increase the share of manufacturing in output. - Further increase trade (noting trade is still relatively small by East Asian standards). - Required policy actions and investments: - Significant investments in infrastructure. - Significant investments in human capital (Tanzania is lagging Asian peers in health and education indicators). ### Tanzania — potential natural gas discovery and risks - If commercial quantities of gas are confirmed in the near future, Tanzania could: - Receive multi-billion dollar foreign investments in the coming years. - Become one of the largest exporters of natural gas in the region by 2020. - Potential macroeconomic and structural risks from a commodity boom: - Real appreciation. - Increase in labor costs. - Crowding-out of investments in other sectors. - Difficulties in containing inefficient public spending. - Inflationary pressures. - Highly profitable extractive sectors absorbing most productive labor while trapping the rest of the economy in low value-added production. - Recommended policy response: - A comprehensive policy and institutional framework for managing natural resource wealth to avoid loss of competitiveness in other tradable sectors and to sustain growth in manufacturing. - Additional reforms to strengthen the business environment and remaining non-tariff barriers. - Harmonize regional norms and investment rules to develop and fully exploit comparative advantages within and outside the region. ### Comparative and contextual notes (excerpts referencing other countries) - Tanzania’s experience is compared to Vietnam as an example where removal of constraints aided diversification. - Figures and time-series comparisons (1996–2011; 1997 base indices; 1991–2011 destination shares) are cited to illustrate sectoral and export shifts. *Source: _030514a - 150. Gold has become the single largest export, accounting for 9¼ percent of GDP in 2011, (https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2014/_030514a.pdf)* ### 169. After the conclusion of the peace agreement in 2001, a massive construction effort ### _030514a - 169. After the conclusion of the peace agreement in 2001, a massive construction effort ### Post-war reconstruction and investment - A massive construction effort was launched after the 2001 peace agreement, financed by oil revenue. - First phase: repairing essential infrastructure and addressing conditions of a large, displaced population. - Mid-2000s onward: a second wave emphasizing housing and transport infrastructure. - These large public investments, together with demand from the oil sector, fueled strong growth in manufacturing, construction, transportation, and (business) service sectors and contributed to modest diversification. ### Sectoral impacts and oil dependence - Annual GDP growth during 2002-08 averaged 15¼ percent. - Three quarters of GDP is concentrated in the capital Luanda. - The public sector dominates the economic sphere. - Oil sector dominance: - Angola’s export composition became increasingly oil-dominated; the share of oil in total goods exports rose to 99 percent in 2011. - Replacement of the United States by China as the largest buyer of Angolan oil coincided with an increasing commercial Chinese presence in oil and construction sectors and a large (and growing) Chinese labor force, which limited domestic spillovers. - The global financial crisis highlighted oil dependence: oil revenue declined, oil revenue flows to the treasury became volatile, the exchange rate depreciated, and a sharp fiscal consolidation and growth slowdown were necessary. ### Agriculture and missed opportunities - Agriculture did not benefit from the post-war economic boom. - Angola continues to import about 90 percent of its food. - Fertile land in the east and center remains largely inaccessible due to the legacy of mining and inadequate transport infrastructure. - Storage facilities are insufficient for large cities; there is virtually no access to finance. - As a result, agriculture remains small and contributes little to economic growth despite large potential. ### Challenges to diversification - Diversifying the economy is critical to reduce vulnerabilities and achieve inclusive and sustainable growth. - The government has shifted policy focus toward structural transformation and diversification. - Required reforms and investments include: - Regulatory and governance reform to improve the investment environment. - Macroeconomic policies aimed at reducing volatility and securing stability. - Investment in transport, power, and water infrastructure to support industrial development and unlock agricultural potential (with implications for employment and poverty reduction). - Investment in health and education to improve skills and labor productivity over time. - Obstacles: - Oil wealth can crowd out other sectors and render them uncompetitive (Dutch disease via real exchange rate appreciation and loss of competitiveness). - Large natural resource sector size may undermine diversification despite general policies. - Deliberate industry-specific support risks intensifying rent-seeking in environments of weak governance. ### Policy options and trade-offs - Broad-based policies likely needed: macroeconomic stability, removal of barriers, public investment, and investment in people and infrastructure. - Deliberate tilting of the playing field (targeted support for specific industries) has mixed international experience: - Some countries (Malaysia) show deliberate action can help diversify. - Others (Vietnam, Tanzania) suggest broader-based policies matter more. - State support has pros and cons: - Can underpin specific industries but may increase export concentration and carry opportunity costs (foregone tax revenue or direct expenditure). - In Angola, policies to tilt the playing field away from oil may be required, but must account for governance-related risks of rent-seeking. ### Key observations and lessons (from the broader case sample) - Entrepreneurs are central to diversification and growth; reducing costs of branching into new activities is important. - Important ingredients supporting entrepreneurship and diversification: - Macroeconomic stability (reduces volatility and hedging needs). - Removing barriers to entry and operation (lowers costs, encourages expansion). - Investment in infrastructure (lowers business costs; public role significant). - Investment in people (education and skills facilitate movement up the quality ladder and foster entrepreneurship). - Natural-resource endowment can support diversification if the “resource curse” is avoided; examples show mixed experiences depending on resource scale and policy choices. *Source: IMF content unit _030514a - 169.* --- _Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2014/_030514a.pdf_