Globalization, Networks, and the Interconnectedness of Europe and Central Asia (ECA) What s at Stake for Inclusive Growth? David Gould The World Bank 25 January 2018 GICA Conference Paris
Why this report? Connectivity or globalization and its relationship to economic growth is often viewed in only one dimension (i.e., trade or FDI) through its impact on the transfer of ideas and technological knowledge (tacit and explicit). o But various connections are often complements Ideas are transmitted through complementary connections o Migration, trade, FDI, internet, finance, transport, etc. Is a more explicit way of examining the why we get Economic Complexity beyond trade To whom you are connected to might be just as important as the type of connection Trade with Germany may be more important than trade with Algeria not only due to the level of technological potential embodied in exports/imports but also the connections of your partner s connections Connectivity being multidimensional implies that shocks in one dimension (say migration) can have adverse affects in other dimensions (say FDI and trade) as well. 2
Key findings Being well-connected in the global network of countries of is important for long-run inclusive economic growth. ECA policies that build regional and global connectivity will stimulate robust growth Because of complementarities in types of connections (trade, FDI, migration, etc.), a balanced connectivity profile maybe more important for growth than being well connected in a single dimension. Deep and comprehensive FTAs are the best way to achieve this Europe and Central Asia have benefited from globalization (in many dimensions) and it would be naive to believe cutting connections in one dimension, or to a particular country, would not have broader implications. 3
What Multidimensional Connectivity Looks Like Individual Network Layers Trade Network FDI Network N-Network Multidimensional Network (collapsed) MDC Network 4
Example: Trade has grown within ECA, although China is taking a larger share of global trade The size of each country-node reflects the total volume of trade. Each node has two outgoing links, which point to the two top export partners of the country. 2000 2014 Blue nodes are Europe and emerging ECA countries 5
Example: ECA s airline connectivity has grown with emergence of low-cost carriers and regional integration The size of each country-node represents the estimated total number of air passengers. Each country node has two outgoing links that point the two largest destination countries 2000 2012 Blue nodes are Europe and emerging ECA countries 6
Example: ECA s intra-regional ICT has not increased relative to the rest of the world Each country node represents the combined value of the estimated incoming and outgoing ICT communication. Two outgoing links point to the two main outgoing partner countries 2002 2010 7
Multidimensional Connectivity Network The size of the node represents the multidimensional connectivity index of each country. The two outgoing links point to the strongest two connections 2000 2014 8
Multidimensional Connectivity and Growth Multidimensional Connectivity The dependent variable in the model is annual GDP growth (%). The PageRank multidimensional connectivity coefficient is standardized to represent the effect of a change of one standard deviation. Overall Growth Multidimensional Connectivity Impact on Growth 0.67*** Efficiency Exponents/Weights of Connectivity Channels Trade Channel Efficiency 0.53 FDI Channel Efficiency 0.37 Migration Channel Efficiency 0.10 ICT Channel Efficiency 0 Airline Channel Efficiency 0 Portfolio Flows 0 Adj-R 2 0.61 9
Thank you 10
ECA s intra-regional migration has increased and is a larger share of global migration Country-node represents the number of foreign born individuals in the country and number of native born living outside the country. Outgoing links show the two largest emigration destinations of each country 2000 2010 Blue nodes are Europe and emerging ECA countries 1 1
ECA s inward FDI has grown, but inflows have increased from outside the region The size of the country-node reflects the total FDI stocks (incoming and outgoing) of the country. Each country has two outgoing links that point to the two main FDI destinations for each country 2003 2012 Blue nodes are Europe and emerging ECA countries 1 2
Shock originating in Germany (10 Percent fall in all types of connectivity) 0-1 -2-3 -4-5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Initial level of per capita connectivity Change in level of connectivity -6 Blue nodes are Europe and emerging ECA countries 1 3
Shock originating in USA (10 Percent fall in all types of connectivity) 0-2 -4-6 -8-10 -12 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Initial level of per capita connectivity Change in level of connectivity -14 Blue nodes are Europe and emerging ECA countries 1 4
Shock originating in Russia (10 Percent fall in all types of connectivity) 0-0.02-0.04-0.06-0.08-0.1 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Initial level of per capita connectivity Change in level of connectivity -0.12-0.14 Blue nodes are Europe and emerging ECA countries 1 5
Most foreign firms in ECA are owned by Germany and the United States Note: sample excludes firms owned by tax havens. Each cell shows the first five (or more, if there is a tie) most common countries of ownership, among the top 10 countries of ownership for each country. For the Russian Federation and Ukraine, the cells show the 5 countries with the biggest ownership shares. 16
Foreign affiliates have better management practices than local firms Source: Own calculations based on WEF and Orbis data 17
Foreign owned and managed firms perform better than local firms Source: Based on regression analysis. Growth period 2010-2013. All of the underlying coefficients are statistically significant. 18
The positive spillovers of well-managed foreign firms seem weaker for small and young firms Regional Spillover Effects Sectoral Spillover Effects Note: Based on regression analysis. Small firms are those with 49 employees or less; young firms are those aged 4 years or less. Each bar is the effect of increasing 1 point the average management scores of foreign firms. The bars in each panel are the baseline effect, the baseline effect plus the interaction term associated with size, and the baseline effect plus the interaction terms associated with size and age simultaneously. 19
Growth Model with Network Effects 20
Classic PageRank Algorithm 21
Multidimensional PageRank Measure (collapsed network) 22