Economic Complexity Index
The Economic Complexity Index (ECI) is a quantitative measure of an economy's productive knowledge and capabilities, derived from the diversity of goods it exports and the ubiquity of those goods across other economies, with higher values indicating greater sophistication and potential for sustained growth.[1] Developed by physicist César Hidalgo and economist Ricardo Hausmann in 2009, the index applies network theory to trade data, assigning complexity scores to products based on the income levels and diversification of exporting countries, then aggregating these for economies while penalizing reliance on common exports.[2] Empirically, the ECI outperforms traditional predictors like education or institutions in forecasting GDP per capita and long-term growth rates, as evidenced by regressions on panel data from over 100 countries spanning decades.[1][3] ![Rank in the Economic Complexity Index, OWID.svg.png][float-right] Published annually through platforms like the Observatory of Economic Complexity and Harvard's Atlas of Economic Complexity, the ECI ranks nations— with Switzerland, Singapore, and Japan typically leading due to exports in high-tech machinery, pharmaceuticals, and precision instruments— and informs policy by highlighting paths to diversification beyond resource dependence.[4] Its causal intuition rests on the accumulation of productive knowhow as a driver of prosperity, where complex economies embody dense networks of specialized skills that enable adaptation to global demands, though critics note potential distortions from misreported trade or services undercounting.[2] Applications extend to subnational regions and projections, where gains in ECI correlate with reduced inequality and resilience, underscoring its role in empirical development economics over ideologically driven narratives.[1]Origins and Theoretical Foundations
Historical Development
The concept of economic complexity underlying the Economic Complexity Index (ECI) originated in research initiated around 2006, focusing on mapping the "product space" to understand how countries transition between exported goods based on relatedness in production capabilities.[3] This work built on network analysis techniques to visualize trade patterns, revealing that economic development follows paths constrained by proximity in this space rather than random diversification. The foundational publication appeared in 2007 as "The Product Space Conditions the Development of Nations" in Science, authored by César A. Hidalgo, Bailey Klinger, A.-L. Barabási, and Ricardo Hausmann, which introduced the product space metric but did not yet formalize the ECI. Hidalgo, then at the MIT Media Lab, and Hausmann, at Harvard's Center for International Development, extended this framework in 2009 with "The Building Blocks of Economic Complexity" in Proceedings of the National Academy of Sciences (PNAS), where they defined the ECI using the method of reflections on a bipartite network of countries and products to quantify a country's productive knowledge beyond mere diversification or ubiquity.[2] This index aggregated export data to measure latent capabilities, positing that higher complexity correlates with sustained growth potential. The ECI gained prominence through the first edition of The Atlas of Economic Complexity, released on October 27, 2011, at Harvard's Global Empowerment Meeting, which visualized ECI rankings and product spaces for over 100 countries using Harmonized System trade data from 1995–2009.[3] Subsequent updates, including the 2013 edition, incorporated refinements like annual computations and expanded datasets, while the Observatory of Economic Complexity (OEC), launched in 2014 by Hidalgo, provided open-access tools for real-time ECI tracking based on the original methodology.[5] These developments emphasized empirical validation through correlations with GDP per capita growth, distinguishing ECI from traditional metrics like export sophistication by incorporating network effects.[2]Core Concepts of Economic Complexity
Economic complexity quantifies the productive knowledge and capabilities embedded within an economy, manifesting in its capacity to produce a diverse set of sophisticated goods and services that require coordinated skills, technologies, and institutions. This perspective posits that wealth generation stems primarily from the accumulation and combination of such knowhow, rather than solely from natural resources or capital inputs, enabling sustained higher incomes and resilience to shocks. Hidalgo and Hausmann argue that a country's productivity resides in the diversity of its nontradable capabilities, with economic complexity serving as a proxy for these latent factors that explain cross-country differences in prosperity.[2][6] Central to the theory are the intertwined measures of diversity—the variety of products an economy exports competitively—and ubiquity—the prevalence of those products across other economies. High-diversity economies tend to specialize in low-ubiquity products, which few countries can produce due to the rare capabilities required, such as advanced engineering or specialized supply chains. In contrast, ubiquitous products, like basic commodities, signal simpler production processes accessible to many nations. This dynamic reveals an economy's sophistication: complex structures export non-routine, knowledge-intensive goods, reflecting deeper pools of collective expertise accumulated over time through path-dependent learning.[2][1] The product space conceptualizes economies as networks of related activities, where products are connected if countries frequently co-export them, illustrating viable paths for diversification. This relatedness principle underscores that productive transformation occurs incrementally, as nations build upon proximate capabilities to enter more complex sectors, avoiding infeasible leaps into unrelated industries. Economic complexity thus frames development as an evolutionary process driven by capability accumulation, with indices like the ECI forecasting outcomes such as income growth—typically 4-7% annualized per standard deviation increase—by identifying gaps between current complexity and realized prosperity.[6][1]Methodology and Computation
Data Inputs and Revealed Comparative Advantage
The Economic Complexity Index (ECI) draws on bilateral trade data, primarily export flows, compiled from the United Nations Commodity Trade Statistics Database (UN Comtrade), which aggregates reported values from national customs authorities across approximately 200 countries and territories.[7] These data encompass merchandise goods classified under the Harmonized System (HS) nomenclature, typically at the six-digit level for granularity, but aggregated to the four-digit HS level for ECI computations to balance detail with statistical robustness, yielding over 1,200 product categories.[1] Export values are expressed in current U.S. dollars, with annual updates reflecting the latest available reporting, though lags in data submission can introduce minor discrepancies for recent years.[8] Services trade data from the International Monetary Fund's Direction of Trade Statistics are occasionally incorporated in extended analyses but excluded from core ECI calculations due to inconsistent global coverage and classification challenges.[8] To construct the foundational bipartite network of countries and products, the revealed comparative advantage (RCA) metric, originally formulated by Béla Balassa in 1965, filters trade data to identify specialized capabilities. For a given country c and product p, RCA is calculated as: \text{RCA}_{c,p} = \frac{X_{c,p} / X_c}{X_{w,p} / X_w} where X_{c,p} is country c's exports of product p, X_c is country c's total exports, X_{w,p} is global exports of product p, and X_w is global total exports.[9] An RCA value exceeding 1 signifies that country c exports product p at a higher relative intensity than the world average, implying a comparative advantage rooted in productive knowledge or capabilities rather than mere scale.[2] This threshold binarizes the trade matrix, setting M_{c,p} = 1 for RCA > 1 (indicating "presence") and 0 otherwise, which mitigates noise from small or non-competitive trade volumes and emphasizes structural economic features over absolute trade sizes.[10] The RCA approach assumes that sustained export specialization reveals underlying productive capacities, as countries tend to export goods aligned with their knowledge-intensive strengths, though it overlooks domestic consumption, informal sectors, and non-tradeable outputs.[11] Empirical validation shows RCA stability over time for diversified economies, with correlations exceeding 0.9 across consecutive years, supporting its use as a proxy for comparative advantage despite critiques of endogeneity from global demand shifts.[2] In ECI methodology, the binarized RCA matrix underpins subsequent iterations to derive country-level complexity scores, prioritizing economies with diverse, non-ubiquitous export baskets.[1]Index Formulation and Algorithms
The Economic Complexity Index (ECI) quantifies the knowledge intensity and productive capabilities of an economy by analyzing the diversity and sophistication of its export basket, derived from international trade data at the Harmonized System (HS) 6-digit level. The computation begins with the construction of a bipartite matrix M, where M_{c,p} = 1 if country c reveals a comparative advantage (RCA) in product p, and 0 otherwise; RCA is defined as RCA_{c,p} = \frac{X_{c,p}/X_c}{X_{w,p}/X_w}, with X_{c,p} denoting exports of product p from country c, X_c total exports of country c, and X_w, X_{w,p} the world equivalents, using data typically from sources like UN Comtrade for years such as 2018–2022 in recent iterations.[1][2] This threshold of RCA > 1 identifies products where a country's export share exceeds the global average, filtering for non-trivial specializations. The core algorithm, termed the Method of Reflections, iteratively estimates the ECI for countries and the complementary Product Complexity Index (PCI) for products to resolve circular dependencies: complex economies produce complex products, and complex products are produced by complex economies. Define country diversity k_c = \sum_p M_{c,p} and product ubiquity k_p = \sum_c M_{c,p}; the row-normalized matrix is \tilde{M}_{c,p} = M_{c,p} / k_c (probability of product p given country c), and column-normalized \tilde{M}_{p,c} = M_{p,c} / k_p. Initialize PCI^{(0)}_p (often as 1 or inversely to ubiquity, e.g., -\log(k_p / N_p) where N_p is total product-country links), then iterate: ECI^{(t)}_c = \sum_p \tilde{M}_{c,p} PCI^{(t-1)}_p , followed by PCI^{(t)}_p = \sum_c \tilde{M}_{p,c} ECI^{(t)}_c , normalizing at each step (subtract mean, divide by standard deviation) until convergence, typically after 10–20 iterations.[1][2][9] This process yields the fixed-point solution equivalent to the leading eigenvector of the country-country similarity matrix \tilde{M} \tilde{M}^T, where similarity reflects co-exported products, capturing higher-order proximities beyond direct overlaps.[9] The resulting ECI values are standardized to have zero mean and unit variance across countries, with higher scores indicating economies diversified into less ubiquitous (rarer) products, interpreted as embodying greater productive knowledge.[1] Computationally, it leverages linear algebra for scalability, as implemented in tools like Python's NetworkX or the OEC's open-source codebase, and has been validated to converge robustly across datasets spanning 1962–2022.[1] Variations in implementation include handling missing data via imputation or restricting to HS codes with sufficient trade volume (e.g., >$1 million annually), and sensitivity analyses confirm stability to such adjustments, though early versions used coarser SITC classifications before standardizing to HS.[2] The algorithm's self-consistent nature avoids arbitrary priors, privileging empirical network structure over subjective weights, though critics note its reliance on export RCA may underweight services or domestic production.[9]Updates and Variations
The Economic Complexity Index (ECI) has undergone periodic updates primarily through refreshed datasets rather than fundamental methodological overhauls, with annual recalculations incorporating the latest bilateral trade flows from sources such as UN Comtrade at the HS6 product classification level.[1] These updates, facilitated by the Observatory of Economic Complexity (OEC) and Harvard Growth Lab's Atlas of Economic Complexity, ensure the index reflects current export structures; for instance, the Atlas version 10.0, released in September 2024, integrated trade data up to 2022 alongside enhanced visualizations like an updated Product Space, though the core ECI algorithm remained unchanged.[12] Such data refreshes have revealed shifts in rankings, such as Singapore maintaining top positions due to sustained diversification in high-tech exports, while resource-dependent economies like those in sub-Saharan Africa show slower complexity gains amid volatile commodity trades.[4] Variations of the ECI extend its framework to non-trade data, broadening its application to measure productive capabilities in innovation and knowledge domains. The OEC computes ECI equivalents using patent filings (ECI technology) and scientific publications (ECI research), which correlate positively with trade-based ECI but highlight discrepancies; for example, nations like the United States rank higher in patent complexity due to R&D intensity, underscoring trade data's limitations in capturing upstream innovation.[13] A multidimensional variant combines trade, patent, and publication metrics into a composite index, explaining up to 60% of variance in future GDP growth and reducing inequality when inclusive policies align with complexity gains, as evidenced in panel regressions across 100+ countries from 2000–2018.[14] Alternative formulations address perceived shortcomings in the original ECI's iterative averaging method, which can amplify noise in sparse networks. The Fitness and Complexity Index (FCI), developed in 2012, employs a binary bipartite network projection with a fitness metric based on product "fitness" (probability of export success) and country complexity, yielding non-negative values and stronger predictive power for growth in developing economies by prioritizing rare, sophisticated products over ubiquity adjustments. Another variant, the Enhanced ECI (ECI+), refines export sophistication by correcting for global value chain distortions and importer biases, applied in studies showing improved correlations with governance quality in emerging markets from 1995–2020.[15] These adaptations, while retaining the diversity-ubiquity paradigm, often outperform the baseline ECI in robustness tests against endogeneity, though empirical debates persist on whether they truly capture causal productive knowledge or merely repackage trade fitness.[16]Empirical Evidence and Predictive Power
Correlations with Economic Outcomes
The Economic Complexity Index (ECI) exhibits a strong positive correlation with gross domestic product (GDP) per capita across countries, reflecting the association between productive sophistication and income levels. In analyses of data from 1998 to 2000, the Pearson correlation coefficient between the logarithm of GDP per capita (purchasing power parity-adjusted) and higher-order measures of economic complexity (derived via the method of reflections) reaches values exceeding 0.7, surpassing simpler diversification metrics like the Herfindahl-Hirschman index.[2][17] This relationship holds after controlling for factors such as natural resource endowments, which can inflate GDP without corresponding productive capabilities.[18] Beyond contemporaneous associations, the ECI demonstrates predictive power for future economic growth. Regressions using 1985–2005 data show that deviations from the expected income level given a country's complexity—where higher complexity implies potential for elevated income—forecast subsequent GDP per capita growth rates over 5-, 10-, and 20-year horizons, with higher-order complexity indices yielding stronger predictions (e.g., standardized coefficients indicating up to 1% additional annual growth per standard deviation increase in complexity, conditional on initial income).[2][17] This outperformance relative to traditional indicators underscores the ECI's emphasis on latent productive knowledge as a driver of sustained development, rather than static resource bases or input factors alone.[19] Empirical extensions link ECI to ancillary outcomes, including innovation proxies. Countries with elevated ECI rankings in the early 2000s subsequently exhibited higher growth in patent filings and scientific publications per capita through the 2010s, suggesting that economic complexity captures capabilities enabling technological advancement and knowledge accumulation.[6] These patterns persist in panel data analyses, where ECI Granger-causes growth increments independent of institutional variables, though causality inferences remain probabilistic given observational data limitations.[20]Superiority Over Traditional Metrics
The Economic Complexity Index (ECI) demonstrates superior predictive power for long-term economic growth compared to traditional metrics such as GDP per capita, which primarily reflect current output levels but fail to capture underlying productive capabilities. In regression analyses spanning 1972–2008, a one-standard-deviation increase in ECI correlates with approximately 1.9% higher annual GDP per capita growth, explaining a substantial portion of cross-country growth variations that GDP per capita alone cannot, as the latter often implies convergence trends unsupported by empirical divergence in complexity-driven economies.[3] When combined with the Complexity Outlook Index (COI), ECI accounts for about 50% of the variance in 10-year GDP per capita growth across panels from 1975–2005, outperforming initial income levels (a proxy for GDP per capita) which contribute less to explanatory power in the same models.[3] Traditional metrics like export sophistication (EXPY) yield lower R² values (0.367) in growth regressions compared to ECI's 0.472, with EXPY losing significance when both are included, indicating ECI's ability to better isolate non-trivial, knowledge-intensive advantages.[3] ECI also surpasses composite indices like the Global Competitiveness Index (GCI) in forecasting growth, with ECI rankings explaining up to 15.5% more variance in 10-year growth rates than GCI rankings.[3] Unlike Human Development Index (HDI) components—such as education years, which add only marginal R² improvements (around 0.01–0.03)—ECI coefficients remain robust (0.011–0.013, p<0.01) in growth models, reflecting its focus on embodied know-how through export diversification and ubiquity rather than averaged inputs or outputs.[3] This structural emphasis enables ECI to predict sustained prosperity in complex economies, avoiding the pitfalls of metrics biased toward resource rents or short-term financial depth, which show insignificant growth correlations when controlling for complexity.[3]Applications in Forecasting and Policy
The Economic Complexity Index (ECI) has demonstrated utility in forecasting long-term economic growth by leveraging historical correlations between a country's productive capabilities and subsequent GDP per capita increases. Regressions from 1978 to 2008 indicate that ECI and the related Complexity Outlook Index (COI) explain approximately 50% of the variance in 10-year growth rates, with a one-standard-deviation increase in ECI associated with an additional 1.9% annual growth after controlling for initial income levels.[3] More recent models incorporate multidimensional ECI variants—drawing from trade, patents, and research outputs—to predict growth through 2032, achieving an adjusted R-squared of 0.306 in calibrations over 1999–2021 data, outperforming trade-only ECI by 4 percentage points.[21] For instance, projections estimate India achieving 4.2% average annual growth and the Philippines 3.8% over 2022–2032, conditional on baseline complexity and income convergence dynamics.[21] ECI's forecasting applications extend to subnational levels and alternative outcomes, such as city population growth or sectoral shifts, by adapting the index to local export or industry data, though predictive accuracy diminishes without robust capability proxies.[3] These models robustly account for endogeneity via instrumental variables like non-neighboring countries' complexity levels, underscoring ECI's emphasis on latent productive knowledge over observable factors like natural resources.[21] In policy formulation, ECI informs strategies for economic diversification by mapping "adjacent possibles"—products or industries proximate in the product space (high relatedness) yet more complex—to build capabilities incrementally.[3] This approach guides targeting of feasible, high-complexity exports, as seen in frameworks prioritizing relatedness-complexity diagrams for investment decisions, such as Shanghai's focus on spark ignition engines or Turkey's analogous shifts.[22] Governments have operationalized this via data platforms like DataMéxico, which supports Mexico's "smart diversification" by revealing capability-aligned opportunities, and similar observatories in Peru and Brazil for export promotion.[22] Policy applications emphasize timing and agents: low-complexity economies prioritize related diversification to avoid lock-in, while leveraging migrants or foreign firms for knowledge transfers, as in Hungary's firm-level analyses.[22] Harvard's Growth Lab has applied ECI in advisory reports, such as Kazakhstan's, recommending capability-building in non-resource sectors to enhance resilience.[23] Broader industrial policies integrate ECI to evaluate incentives for risky investments, focusing on green technologies or innovation networks where relatedness accelerates structural change.[24][22] Such uses highlight ECI's causal insight that productive knowledge, rather than institutions alone, drives sustained development, though implementation requires complementary investments in education and infrastructure.[3]Criticisms and Limitations
Methodological Shortcomings
The Economic Complexity Index (ECI) relies exclusively on gross export data, which fails to account for global value chains where intermediate inputs cross borders multiple times, potentially overstating or understating a country's true productive capabilities.[10] This approach also neglects non-tradable sectors such as services, construction, and domestic production, underestimating complexity in economies oriented toward internal markets or service-based activities.[25] For instance, service complexity measures exceed those of goods in many cases, yet ECI excludes them entirely due to data limitations in harmonized trade classifications.[25] A core methodological flaw stems from the binary application of Revealed Comparative Advantage (RCA), using a fixed threshold of RCA ≥ 1 to determine whether a country "exports" a product, which introduces discretization noise and arbitrary cutoffs sensitive to economy size and market fluctuations.[10] This binarization discards granular export value information, reducing the index's precision and making it vulnerable to outliers in trade data.[10] Furthermore, the underlying Method of Reflections algorithm lacks a formal theoretical foundation for defining "economic complexity," relying instead on iterative eigenvector approximations that correlate empirically with income but may proxy institutional or historical factors rather than causal productive knowledge.[11][26] Variant methodologies, such as the Fitness-Complexity approach, yield divergent country rankings from the ECI's Method of Reflections, with significant scatter in outcomes due to differences in handling diversification, ubiquity, and non-linearity—undermining the robustness of ECI as a universal measure.[16] The algorithm assumes capabilities are fully embodied in national exports, ignoring offshoring and vertical integration in supply chains, which distorts assessments for intermediate goods-heavy economies.[25] These inconsistencies highlight a broader absence of convergence criteria or validation against micro-level firm data, limiting the index's applicability beyond trade-focused analyses.[16][11]Empirical Debates and Conflicting Findings
Different methodologies for assessing economic complexity, such as the Method of Reflections underlying the Economic Complexity Index (ECI) and the Fitness and Complexity (FC) approach, produce substantially divergent country rankings, with wide scatters observed in comparative evaluations.[16] The ECI's linear aggregation loses information on export diversification and ubiquity, while FC's non-linear method adjusts product quality based on exporter competitiveness, leading to inconsistencies that question the ECI's robustness for cross-country comparisons.[16] These methodological contrasts have fueled debates on whether the ECI accurately captures underlying productive capabilities or merely reflects data artifacts.[16] Empirical critiques further highlight deficiencies, including biases from using gross exports, which overestimate complexity in developed countries positioned higher in global value chains by ignoring intermediate inputs.[27] [11] Studies report low predictive power for economic growth when revisiting ECI correlations, attributing this to overreliance on export composition without accounting for domestic value added or confounding factors like institutional quality.[28] Additionally, analyses of output volatility reveal ambiguous relationships, where higher ECI scores do not consistently reduce macroeconomic fluctuations across panels of countries from 1960 to 2018.[29] While proponents cite ECI's orthogonality to simpler diversification metrics as evidence of added explanatory value for GDP per capita and growth, skeptics contend these correlations weaken under alternative specifications or regional data, suggesting limited causal insight beyond traditional predictors.[9] Such conflicting findings underscore ongoing tensions between the index's intuitive appeal and empirical validation, prompting calls for frameworks integrating total factor productivity and supervised algorithms to enhance forecasting stability.[11]Alternative Approaches
The Economic Fitness and Complexity framework, developed by Tacchella et al. in 2012, offers an alternative to the ECI by employing a non-linear, iterative algorithm that calculates a country's fitness as the sum of probabilities associated with exporting specific products, where product quality is refined through successive iterations emphasizing rarity and diversification.[30] This method penalizes reliance on ubiquitous goods more severely than the ECI's linear method of reflections, resulting in rankings that correlate highly with ECI at the top and bottom (e.g., advanced economies like Japan and commodity-dependent nations like those in sub-Saharan Africa) but diverge significantly for middle-income countries, with rank correlations around 0.8-0.9 in empirical comparisons.[31] Proponents argue it better captures underlying productive capabilities by avoiding the information loss from ECI's averaging of product complexities, though it requires careful initialization to prevent convergence issues in computations.[16] A structural approach, outlined by Everett et al. in 2019, derives country rankings from a multi-product Eaton-Kortum trade model, estimating latent productivities via fixed-effects regressions (using OLS or Poisson pseudo-maximum likelihood) on bilateral trade flows at the HS-4 digit level for 127 countries in 2016, followed by eigenvector decomposition of a derived country-productivity similarity matrix.[32] Unlike the ECI's reliance on binary revealed comparative advantage thresholds, this continuous measure incorporates trade frictions and comparative statics, yielding top rankings for Japan, South Korea, and Switzerland, and bottom for Yemen, Sudan, and Malawi, with a 0.96 correlation to ECI rankings.[32] It demonstrates robustness across estimation techniques, as PPML and OLS rankings correlate above 0.995, but requires stronger parametric assumptions about trade elasticities. To address discrepancies between reflection-based (like ECI) and fitness-based methods, Balland et al. in 2020 proposed the GENEPY index, a multidimensional metric recasting both into a unified eigen-problem framework using a symmetric proximity matrix from export data, where complexity emerges from the quadratic form of leading eigenvectors weighted by eigenvalues.[16] This hybrid preserves diversification signals from fitness approaches while retaining the linearity of reflections, tracking trajectories in a two-dimensional space that reveals path dependencies in economic upgrading, such as sustained complexity gains in East Asian economies from 1960-2017.[16] Empirical tests show GENEPY's predictive power for growth rivals or exceeds individual methods, though it demands higher-dimensional data processing.[16] Other variants extend beyond trade data, such as multidimensional complexity indices combining exports with patent filings to proxy innovative capabilities, explaining up to 40% of variance in future growth and inequality reductions across 110 countries from 1995-2019.[14] These alternatives generally affirm ECI's core insights on complexity's role in development but underscore sensitivities to binarization, linearity, and data granularity, prompting ongoing refinements for policy applications.[33]Rankings, Tools, and Visualizations
Country and Regional Rankings
The Economic Complexity Index (ECI) ranks economies according to the diversity and sophistication of their export profiles, with higher values signaling greater productive capabilities. In 2023 data, Singapore topped the rankings at 2.52, driven by exports in electronics, chemicals, and precision instruments that reflect deep technological know-how. Switzerland ranked second at 2.51, supported by pharmaceuticals, machinery, and watches requiring specialized skills. Japan placed third with 2.43, leveraging automobiles, semiconductors, and robotics.[4] East Asian economies dominate the upper echelons, with Taiwan (2.24) and South Korea (2.23) in fourth and fifth, their rankings underpinned by integrated high-value manufacturing clusters in information technology and shipbuilding. European countries follow closely, including Germany at sixth (2.01) via engineered goods and vehicles, and the United Kingdom at seventh (1.81) through aerospace and financial services-linked exports. These rankings correlate with sustained GDP growth, as complex economies export fewer but higher-value goods less replicable elsewhere.[4]| Rank | Country | ECI Score |
|---|---|---|
| 1 | Singapore | 2.52 |
| 2 | Switzerland | 2.51 |
| 3 | Japan | 2.43 |
| 4 | Taiwan | 2.24 |
| 5 | South Korea | 2.23 |
| 6 | Germany | 2.01 |
| 7 | United Kingdom | 1.81 |
| 8 | Ireland | 1.72 |
| 9 | Slovenia | 1.69 |
| 10 | Austria | 1.67 |