Productive and Technological Capabilities and Export Dynamics in Developing Countries

Size: px
Start display at page:

Download "Productive and Technological Capabilities and Export Dynamics in Developing Countries"

Transcription

1 Productive and Technological Capabilities and Export Dynamics in Developing Countries Sebastian Vergara M. Department of Economic and Social Affairs United Nations LINK Meeting ECLAC, United Nations Santiago, Chile, September 2018

2 Motivation - Capabilities do matter Productive and Technological capabilities Export Dynamics Growth & Development Capabilities are highlighted in several strains of the literature: Structuralist tradition: structural change, diversification and specialization in export markets Schumpeterian tradition: technological efforts drive the process of creative destruction and market and trade dynamics Technology and trade: asymmetries in technological capabilities are key factors in explaining trade flows

3 but how capabilities matter is less clear Level of exports of a steel exporter in Turkey is 1.5 times larger than a steel exporter in Mexico, on average Bangladesh exporters of apparel and clothing export to more than 4 destinations, while in Pakistan to only 2 destinations Mexican exporters of machinery & equipment exports more than 6 products on average (6-digit HS classification), while Thailand exporters only 4 products Differences are associated to level of development, size of the economy, openness, size of manufacturing sector and commodity dependency... Do productive/technological capabilities explain these differences? How? Across which export dimensions?

4 Database: 40 developing countries Exporters Dynamics Database (World Bank) Data for 40 developing countries covering the universe of export transactions. Aggregated info. at sectoral level (HS 2-digit). Averages of: Exports per exporter and per entrant Unit prices per exporter Number of exporters (total and per product) Number of products per exporter Number of destinations per exporter Country-sector-year data. Data Capabilities 1.- Proxy for productive capabilities: Economic Complexity Index (ECI) ECI measure country s multiplicity of productive knowledge based on i) diversity of products; ii) ubiquity of its products (based on number of countries that export each product). ECI captures info on what type of capabilities countries are competitive in. 2.- Proxy for technological capabilities: R&D/GDP Key input for knowledge creation & absorption, product/process innovations.

5 What are the questions? Economic complexity Number of exporters (total/per product) Level of Exports per exporter and per entrant Unit prices per exporter product quality R&D Investments Number of products per exporter Number of destinations per exporter Do countries with more productive capabilities have more and larger exporters? Do these exporters benefit from higher unit prices? Are exporters from countries with more technological capabilities more diversified in products and destinations?

6 R&D over GDP Correlation ECI and R&D investments Economic Complexity and R&D investments, China 1.5 Malaysia Kenya 0.5 Mexico Nicaragua y = 0.26x Tajikistan Madagascar Economic Complexity Index Source: author s own elaboration based on data from the WDI and MIT Observatory on Economic Complexity.

7 Capabilities and level of development Economic Complexity and GDP per capita, 2015 R&D investments and GDP per capita, 2015 Economic Complexity Malaysia Mexico Thailand Chile Kuwait Qatar -1.0 R&D over GDP China Malaysia Brazil -1.5 Guinea Ethiopia 0.5 Mozambique Colombia Chile Kuwait -2.0 Mali GDP per capita (in logs) Madagascar Kyrgistan GDP per capita (in logs) Source: author s own elaboration based on data from the WDI and the Observatory of Economic Complexity, MIT Media Lab.

8 Empirical approach

9 (1) Productive capabilities: ECI Export dimension ijt = i + δ t + θeci jt + βx jt + ε ijt (i is sector; j is country; t is year) i) Number of exporters, total and per product ii) Exports per exporter and per entrant iii) Unit prices per exporter Sectoral and Year fixed effects GDP GDP per capita Trade over GDP Manufacturing sector/gdp Commodity-dependent economies Methodology: OLS. Robust standard errors adjusted by clustering at country level.

10 (2) Technological capabilities: R&D Export dimension ijt = i + δ t + θr&d jt + βx jt + R&D jt HighTech i + ε ijt (i is sector; j is country; t is year) HighTech it : dummy - takes value 1 if sector is R&D intensive. Sectoral and Year fixed effects i) Number products per exporter ii) Number of destinations per exporter GDP GDP per capita Trade over GDP Manufacturing sector/gdp Dummy commodity-dependent economies Methodology: OLS. Robust standard errors adjusted by clustering at country level.

11 Results Productive capacities (ECI)

12 Productive capabilities and number of exporters Number of exporters (1) Number of exporters (2) Number of exporters per product (3) Number of exporters per product (4) ECI (2.60)** (1.70)* (2.17)** (1.73)* GDP (8.28)*** (5.74)*** (7.36)*** (6.40)*** GDP per capita (0.029) (-0.77) Trade over GDP (-0.32) (0.47) Manufacturing Sector (1.49) (1.27) Commodity-Dependent (-1.20) (-1.60) Sectoral Dummies Yes Yes Yes Yes Year Dummies Yes Yes Yes Yes R-squared Number of countries Observations 28,921 27,931 27,634 26,707 ECI positively correlated with number of exporters Within sectors, countries with more productive capacities have more exporters, total & per product. As expected, size of the economy is also relevant.

13 Productive capabilities and level of exports Exports per exporter (1) Exports per exporter (2) Exports per entrant (3) Exports per entrant (4) ECI (1.88)* (1.88)* (1.74)* (2.29)** GDP (5.56)*** (5.36)*** (4.24)*** (5.90)*** GDP per capita (0.79) (-2.64)** Trade over GDP (2.31)** (3.66)** Manufacturing Sector (2.49)** (0.09) Commodity-Dependent (3.58)** (1.41) Sectoral Dummies Yes Yes Yes Yes Year Dummies Yes Yes Yes Yes R-squared Number of countries Observations 27,634 26,707 24,195 23,435 ECI is positive and significant within sectors, countries with more capabilities have larger exporters and entrants GDP per capita negatively associated with exports per entrant. The higher the GDPpc, new exporters able to export with lower size (lower trade costs )

14 Productive capabilities and unit prices Unit prices per exporter (1) Unit prices per exporter (2) ECI (4.15)*** (2.18)** GDP (1.81)* (1.57) GDP per capita (2.87)** Trade over GDP (0.90) Manufacturing Sector (-0.41) Commodity-Dependent (0.64) Sectoral Dummies Yes Yes Year Dummies Yes Yes AIC / R-squared Number of countries Observations 21,543 20,616 ECI and level of development are positively correlated with unit prices. Unit prices as a proxy of product quality.

15 Results Technological Capabilities (R&D investments)

16 Technological capabilities and diversification Products per exporter (1) Products per exporter (2) Destinations per exporter (3) Destinations per exporter (4) R&D (2.05)** (2.09)** (1.57) (4.14)*** GDP (2.22)** (2.35)** (2.01)* (0.23) GDP per capita (1.08) (4.17)*** Trade over GDP (2.39)** (1.56) Manufacturing Sector (2.85)** (3.69)*** Commodity-Dependent (1.14) (3.83)*** R&D * High-tech sector (1.20) (2.95)** Sectoral Dummies Yes Yes Yes Yes Year Dummies Yes Yes Yes Yes AIC / R-squared Number of countries Observations 13,107 12,674 13,107 12,674 R&D is correlated with products and destination per exporter Within sectors, countries with more R&D tend to have more products & destinations per exporter. Correlation between R&D and destinations per exporters in high-tech sectors.

17 Are the results driven by sample selection? Robustness checks. The database contains different amount of observations per country (unbalanced sample). Restricted sample 1: using same number of obs. per country (~38 countries) Balanced sample. Restricted sample 2: using countries with at least 500 observations per country (dropping 20% of database) Expanded sample: expanding data to include zeros (not all countries export in every sector) (data expands by 5%). Main results remain changes that do not affect the conclusions.

18 Conclusions Countries with more productive capabilities do have more and larger exporters (intensive and extensive margin of exports) Exporters from countries with more technological capabilities do have higher unit prices for their products, and exporters are more diversified in products and destinations, especially in high-tech Contribution and Implications Comparison between developing countries Diversification Resilience to trade shocks Number/Size of exporters/product quality Medium term and development implications

19 Thanks