Project: FREE policy brief

This policy paper examines trade diversification, export sophistication, and economic complexity in the South Caucasus and Central Asia during 2019–2024. Using detailed product-level trade data, it assesses how concentrated or diversified countries’ exports and imports are, as well as changes in the sophistication of the products they export. Evidence from the Atlas of Economic Complexity is also used to evaluate diversification opportunities based on countries’ productive capabilities. 

The results reveal substantial heterogeneity across the region. Georgia and Kazakhstan maintain relatively diversified export structures, while Armenia and Azerbaijan exhibit increasing export concentration. Export sophistication improves modestly in several countries, particularly Armenia and Uzbekistan. Overall, the findings suggest gradual but uneven structural transformation across the region, with diversification into more complex export sectors remaining limited.

Introduction

International trade plays a central role in shaping economic growth and macroeconomic stability in the South Caucasus and Central Asia (CCA). The economies of the region are highly open, with trade flows accounting for a large share of GDP in most countries. This strong integration into global markets creates important opportunities for growth, but it also exposes these economies to fluctuations in global demand, commodity prices, and international supply chains, potentially with drastic consequences.

This exposure has become particularly concerning in the context of the global economic environment since 2019. A series of major shocks—including the COVID-19 pandemic, Russia’s invasion of Ukraine, and ongoing conflicts in the Middle East—have disrupted global supply chains, energy markets, and transport corridors. At the same time, geopolitical tensions and shifting industrial and trade policies have increased global policy uncertainty. Frequent changes in tariff policies and strategic trade measures by major economies, including the United States, have further contributed to an increasingly uncertain global trading environment.

In this context, the resilience of national economies depends not only on the scale of trade but also on its structure. Countries with concentrated export baskets or strong dependence on a small number of trading partners are typically more vulnerable to external shocks. By contrast, economies with diversified exports and greater participation in higher-value production tend to be more resilient and better positioned for long-term growth. Three related concepts—trade diversification, product sophistication, and economic complexity—provide useful tools for evaluating these structural characteristics. Diversification captures the breadth of the export basket; product sophistication reflects the income and knowledge intensity of exported goods; and economic complexity reflects the broader productive capabilities that underpin them.

This research brief examines the trade structures of the South Caucasus and Central Asian countries through these three dimensions. Using detailed product-level international trade data, the analysis evaluates export and import diversification, the income and technological content of export baskets, and the broader productive capabilities reflected in economic complexity indicators. By comparing patterns across countries and over time, the brief provides new insights into how the region’s economies are positioned to navigate an increasingly uncertain global trading environment.

Stylized Facts: External Balances and Trade Structure

Recent data highlight two closely related characteristics of South Caucasus and Central Asian economies: substantial variation in external balances and strong exposure to international trade. Current account positions differ significantly across the region and fluctuate over time, reflecting differences in export structures, commodity dependence, and import demand (Figure 1). Resource-rich economies such as Azerbaijan, Kazakhstan, and Turkmenistan periodically record sizable surpluses driven largely by oil and natural gas exports, while several other economies experience persistent deficits associated with narrower export bases and higher reliance on imports. Particularly large deficits were observed in the Kyrgyz Republic during 2022–2023, illustrating the sensitivity of smaller economies to shifts in trade flows and external demand. These dynamics are closely linked to the high degree of trade openness observed across the region: smaller economies such as Georgia, Armenia, and the Kyrgyz Republic exhibit particularly high trade-to-GDP ratios, while larger economies such as Kazakhstan and Uzbekistan show somewhat lower—but still substantial—levels of trade exposure.

Figure 1. Current Account as a percentage share of GDP (2019-2025)

Source: Authors’ calculations, IMF

Figure 2 summarizes the geographic composition of trade across the region and highlights the continued importance of a relatively small number of external partners. The European Union, Russia, China, and other CIS economies dominate both export destinations and import sources. Comparing 2019 and 2024, several broad regional shifts emerge. On the export side, the European Union increased its relative importance as a destination for many exports—particularly energy and resource-based products; exports directed to Russia and other CIS markets also grew in importance after 2022. In turn, the share going to the residual category of other countries fell substantially. On the import side, Russia and China strengthened their positions as key suppliers across much of the region, reflecting geographic proximity, established transport corridors, and China’s growing role in regional trade. Imports from the European Union remained important, especially for machinery, equipment, and higher-value manufactured goods.

Overall, despite some adjustments between 2019 and 2024, the region’s trade patterns remain concentrated among a relatively small group of partners. This concentration increases exposure to destination-specific shocks and may weaken trade resilience. It is therefore important to assess not only how diversified exports are, but also how sophisticated and capability-intensive they are, since these characteristics affect an economy’s ability to adapt and redirect trade over time.

Figure 2. Geographic Structure of Regional Exports and Imports for South Caucasus and Central Asian economies, 2019 vs. 2024

Source: Authors’ Calculation, UN Comtrade. Note: Shares of major partner groups (European Union, Russia, China, CIS, and other countries) in total exports and imports of South Caucasus and Central Asian economies.

Methodology

To analyse the structure and evolution of trade patterns, the study employs four complementary indicators: the Herfindahl–Hirschman Index (HHI), the Theil index, the export sophistication indicators PRODY and EXPY, and the Economic Complexity Index (ECI). These indicators capture different aspects of trade structures, including concentration, diversification, technological sophistication, and productive capabilities embedded in economies.

Trade concentration is first evaluated using the Herfindahl–Hirschman Index (HHI). The index measures the extent to which a country’s exports or imports are concentrated across products. In this study, the index is calculated at the HS-4 product level, allowing for a detailed assessment of trade structures.

The Herfindahl–Hirschman Index is defined as:
\[
HHI_i^X = \sum_{k=1}^{N} \left(s_{ik}^X\right)^2
\]
\[
HHI_i^M = \sum_{k=1}^{N} \left(s_{ik}^M\right)^2
\]

where
\[
s_{ik}^X = \frac{X_{ik}}{X_i}
\]
\[
s_{ik}^M = \frac{M_{ik}}{M_i}
\]

Here

• $s_{ik}^X$ = the share of the product in country ’s total exports
• $s_{ik}^M$ = the share of the product in country ’s total imports
• $X_{ik}$ denotes exports of the product  by country
• $M_{ik}$ denotes imports of the product  by country
• $X_i$ denotes the total exports of country
• $M_i$ denotes the total imports of country

The index ranges between 0 and 1. Values close to zero indicate highly diversified trade structures, while values approaching one suggest strong concentration in a limited number of products.

The HHI provides a simple summary of whether trade is concentrated in a small number of products or partners. To complement this, the analysis also uses the Theil index, which captures how unevenly trade is distributed across all destinations or product categories. This distinction matters because similar levels of overall concentration can mask different underlying structures. In plain terms, the Theil index compares the observed trade distribution with a benchmark of equal shares across all categories: a value of zero indicates perfect equality, and the more uneven the distribution, the higher the index. As a result, comparing the two indicators allows a more nuanced assessment of whether concentration changes reflect dominance by a few categories or broader structural shifts in trade patterns.

Unlike the HHI, the Theil index can also be decomposed into within-group and between-group components, which helps identify the sources of concentration. The Theil index for exports and imports is defined as:
\[
T_i^X = \sum_{k=1}^{N} s_{ik}^X \ln\left(\frac{s_{ik}^X}{\bar{s}}\right)
\]
\[
T_i^M = \sum_{k=1}^{N} s_{ik}^M \ln\left(\frac{s_{ik}^M}{\bar{s}}\right)
\]
where
\[\quad \bar{s} = \frac{1}{N}
\]
and N represents the total number of HS-4 products, so that the terms in brackets measure how far the actual share allocation is from the equal share one. Higher values of the Theil index indicate greater concentration of trade across products, while lower values indicate higher diversification. Compared with the HHI, the Theil index has the advantage of being decomposable into within-sector and between-sector components. That is, the HHI can be broken down into group-specific contributions, but it does not provide the same standard additive decomposition with equally clear interpretation. allowing a more detailed examination of diversification patterns. The policy paper assesses export and import concentration using the HHI and Theil indices not only by product categories but also by trading partner countries (in the formulas above, products are replaced by countries).

While the HHI and Theil indices measure concentration and diversification, they do not say much about the type of goods a country exports. To capture this dimension, the analysis uses the PRODY and EXPY indicators introduced by Hausmann, Hwang, and Rodrik (2007). These indicators assess the sophistication of a country’s export bundle, inferring it from the characteristics of the countries that export the respective products. In particular, a product receives a higher PRODY value when it is exported more intensively by higher-income economies. A country’s EXPY then summarizes the sophistication of its overall export basket by taking a weighted average of the PRODY values of the goods it exports.

More specifically, the PRODY index measures the income content of a product and is calculated as the weighted average of the GDP per capita of countries exporting that product:

\[
PRODY_k = \sum_{c} \theta_{ck} \, Y_c
\]

where

\[
\theta_{ck} = \frac{\frac{X_{ck}}{X_c}}{\sum_{c’} \left(\frac{X_{c’k}}{X_{c’}}\right)}
\]

Here

• $Y_c$ denotes GDP per capita of country $c$,
• $X_{ck}$ denotes exports of product $k$ by country $c$,
• $X_c$ denotes total exports of country $c$,

In plain terms, PRODY asks whether a product is typically associated with richer or poorer exporters.

The PRODY calculation gives more weight to countries for which a product is relatively important in the export basket. This prevents the measure from being driven by very small or incidental exports. As a result, a product receives a high PRODY score when it is a meaningful export to richer economies, not merely when it appears in their trade data.

The weights $\theta_{ck}$ capture the relative importance of the product k  in each country’s export basket. Products exported primarily by high-income economies, therefore, receive higher PRODY values.

Using the PRODY values of individual products, the sophistication of a country’s export basket is measured using the EXPY index:

\[
EXPY_i = \sum_{k} \left(\frac{X_{ik}}{X_i}\right) PRODY_k
\]

EXPY applies the same logic at the country level: it shows whether a country’s export basket is tilted toward products that are more commonly exported by higher-income economies. Higher EXPY values, therefore, suggest a more sophisticated export structure – they indicate that the country exports goods that are typically produced by higher-income economies.

Unlike the HHI, PRODY and EXPY do not lie between 0 and 1. Their values are expressed on the scale of the underlying income measure used in the data, so they are most informative in comparative terms across countries and over time. Also, empirical applications frequently use the natural logarithm of EXPY. This transformation reduces skewness and facilitates interpretation in regression analysis.

Finally, to capture the deeper productive capabilities embedded in economies, the analysis incorporates the Economic Complexity Index (ECI) developed by the Harvard Growth Lab and published in the Atlas of Economic Complexity. The ECI measures the knowledge intensity of an economy by combining information on the diversity of products a country exports and the ubiquity of those products across countries.

The calculation begins with the revealed comparative advantage (RCA) indicator:
\[
RCA_{ck} = \frac{\dfrac{X_{ck}}{X_c}}{\dfrac{\sum_{c} X_{ck}}{\sum_{c} X_c}}
\]
where

• $X_{ck}$ denotes exports of product $k$ by country $c$,
• $X_c$ denotes total exports of country $c$.

Countries are considered competitive exporters of product k if their revealed comparative advantage in that product exceeds 1. Based on this country–product matrix relationship, economic complexity is inferred from two simple ideas: diversity and ubiquity. Diversity refers to the number of different products a country can export competitively. Ubiquity refers to how many countries can export a given product competitively. Economies tend to be ranked as more complex (have a higher value of ECI) when they export a broad range of products that relatively few other countries can produce, because this indicates a deeper and more versatile set of productive capabilities.

Taken together, these indicators provide a comprehensive framework for analysing trade structures. The HHI and Theil indices measure trade concentration and diversification at the HS-4 product level for both exports and imports (as well as concentration by countries), the PRODY and EXPY indicators capture the income sophistication of export baskets, and the Economic Complexity Index reflects the underlying productive capabilities of national economies.

Results

The HHI results reveal significant cross-country differences in export diversification. Georgia and Kazakhstan consistently exhibit the lowest export concentration by destination country across the period, with HHI values remaining below 0.15. Although both countries experience a gradual increase in concentration over time, their exports remain comparatively diversified across destination countries relative to the rest of the region.

In contrast, Armenia and Azerbaijan show a noticeable increase in export concentration by destination country after 2021. Armenia’s export HHI rises sharply and remains close to the upper benchmark threshold by 2024, suggesting that exports became increasingly reliant on a smaller set of countries. Azerbaijan also shows a temporary increase in export concentration around 2022, followed by a modest decline by 2024, indicating partial normalization after the peak of external shocks.

The Kyrgyz Republic and Uzbekistan exhibit persistently higher average export concentration by destination country than other countries in the region. Kyrgyzstan reaches particularly high levels during the early pandemic years and remains relatively concentrated thereafter. Uzbekistan also maintains a relatively high concentration, although its export structure shows some signs of gradual diversification toward the end of the period. Tajikistan remains in the intermediate range, with export concentration by country relatively stable across years.

Import concentration patterns differ from those of exports. Several countries maintain relatively diversified import structures by source country throughout the period. Georgia and Azerbaijan show consistently low import HHI values, indicating broad import structures. However, for some other countries in the region, import concentration increases sharply during the shock period. Kazakhstan experiences a substantial increase during 2020–2022, followed by a return to lower levels in subsequent years. Armenia also records a sharp increase in import concentration in 2024, suggesting increased reliance on a narrower set of partner countries. Kyrgyzstan shows a gradual increase toward the end of the sample period.

Figure 3. HHI of Export by Country, 2019–2024

Source: Author’s Calculation, UN Comtrade

Figure 4. HHI of Import by Country, 2019–2024

Source: Author’s Calculation, UN Comtrade

Examining concentration at the HS4 product level provides additional insight into the structure of trade baskets. Changes in overall HHI may arise either because the product distribution becomes more uneven or because a small number of product categories become temporarily dominant.

The HS4-product level export results (Figure 5) reveal substantial cross-country variation in product concentration. Azerbaijan remains the most concentrated exporter throughout the period. Although the figure shows some decline in concentration in the earlier years, this change is not sustained, and the country’s export basket remains heavily concentrated in a narrow set of products. Kazakhstan also shows a relatively high concentration, although the decline after 2022 suggests some gradual diversification. In contrast, Georgia and Tajikistan maintain consistently low HHI values, indicating relatively diversified export baskets across HS4 product categories. Armenia and Uzbekistan remain in the intermediate range, although Armenia shows an increase in concentration in 2024.

Import concentration at the HS4 product level (Figure 6) remains generally lower than export concentration but exhibits greater volatility across countries. Most economies maintain relatively diversified import baskets, with HHI values typically below 0.05–0.06. However, several temporary spikes are visible. Armenia records a sharp increase in import concentration in 2024, suggesting growing reliance on a narrower set of imported goods. Kyrgyzstan experiences a pronounced spike in 2023, while Georgia shows a moderate increase during 2022–2023, then stabilizes. These fluctuations likely reflect temporary supply disruptions, shifts in trade routes, or changes in import demand during periods of economic and geopolitical shocks.

Figure 5. HHI of Export by HS4 Product Categories, 2019–2024

Source: Author’s Calculation, UN Comtrade

Figure 6. HHI of Import by HS4 Product Categories, 2019–2024

Source: Author’s Calculation, UN Comtrade

The country-level Theil index largely reinforces the message from the HHI analysis: across the region, recent changes in export concentration have been driven mainly by shifts in the distribution of exports across destination markets rather than by a restructuring of export baskets.

Armenia shows the clearest increase in geographic concentration, while Azerbaijan also remains relatively concentrated despite some normalization after the 2022 spike. Georgia remains the most geographically diversified case, and Kazakhstan, Kyrgyzstan, and Uzbekistan show only moderate changes over time. Overall, the Theil results add nuance rather than overturning the HHI findings: they suggest that the main source of recent concentration has been unevenness across partner countries, not a uniform narrowing of export structures across all economies.

At the product level, the Theil index points to a more nuanced picture. In several countries, product-level inequality declines or remains moderate even when destination-country concentration rises, suggesting that geographic concentration and product concentration do not always move together. This is especially important for interpretation: an economy may become more dependent on a smaller set of trading partners while still maintaining or even broadening the composition of its export basket. Azerbaijan remains the clearest case of persistently high product concentration, whereas Georgia continues to display a relatively diversified product structure.

In cases where the Theil and HHI measures differ somewhat, the gap likely reflects that the HHI is more sensitive to dominant categories, whereas the Theil index captures unevenness across the full distribution of trade shares.

The product-level Theil index (Figure 7) also provides additional insights into the composition of export baskets. Armenia, Kazakhstan, and Uzbekistan show noticeable declines in product-level inequality between 2019 and 2024, suggesting some diversification across product categories despite rising geographic concentration. This pattern indicates that while exports may increasingly rely on fewer destination countries, the underlying product composition has broadened.

In contrast, Azerbaijan maintains a relatively high product concentration, which is fully consistent with the HS4-product level HHI results showing the highest export concentration across product categories in the region. Georgia shows a slight increase in product-level inequality, although overall concentration remains relatively low compared to most other countries, confirming the diversified structure observed in the HHI product-level analysis.

Overall, the Theil index results reinforce the conclusions drawn from the HHI analysis while providing additional insight into the drivers of concentration. The evidence suggests that recent changes in trade structures across the South Caucasus and Central Asia are driven primarily by shifts in geographic export patterns rather than by widespread narrowing of product specialization. In several countries, product diversification appears to be improving even as exports become more concentrated across trading partners.

Figure 7. Theil Index by Country and Product Categories, 2019 and 2024

Source: Author’s Calculation, UN Comtrade. Note: Higher values indicate greater concentration in the distribution of products.

Export sophistication is measured by the EXPY index, with higher values indicating that a country exports products typically produced by higher-income economies.

Between 2019 and 2024, export sophistication increases for most countries in the region, although the magnitude of change varies (Figure 8). Armenia shows the largest improvement in EXPY, suggesting a shift toward higher-value exports. Uzbekistan and the Kyrgyz Republic also show moderate increases in export sophistication. In contrast, Azerbaijan, Georgia, and Kazakhstan experience only modest changes, indicating relatively stable export structures over the period.

Importantly, increases in export sophistication should be interpreted alongside changes in concentration indicators. When EXPY increases while export concentration remains low or declines, the improvement reflects broader structural upgrading. However, when increases in EXPY coincide with rising concentration, the shift may reflect specialization in a smaller number of higher-value products rather than broad-based diversification.

Figure 8. Export Sophistication (EXPY), 2019 and 2024

Source: Author’s Calculation, UN Comtrade. Note: Higher values indicate a more sophisticated export basket.

The previous indicators evaluate diversification and sophistication based on observed trade patterns. An additional perspective on structural transformation can be obtained by examining future diversification opportunities, using the feasibility analysis derived from the Atlas of Economic Complexity Index (ECI) developed by the Growth Lab at Harvard University (see Figure 9). This framework maps potential export opportunities based on the relationship between product sophistication and proximity to existing productive capabilities.

Figure 9. Economic Complexity Index (ECI), 2012- 2024

Source: Growth Lab at Harvard University

Across the South Caucasus and Central Asia, the feasibility analysis reveals substantial heterogeneity in the pace and depth of structural transformation. Armenia, Kazakhstan, and Uzbekistan show the most pronounced improvements in economic complexity over time, suggesting that a growing number of technologically more sophisticated products are becoming feasible given existing productive capabilities. This pattern indicates a widening diversification frontier and reflects the accumulation of capabilities that can support expansion into more complex sectors.

These findings are broadly consistent with the earlier results on export sophistication (EXPY), which also show noticeable improvements in Armenia and moderate gains in Uzbekistan and Kazakhstan. At the same time, the concentration indicators provide an important qualification. While Armenia shows rising export sophistication, the HHI and Theil indices indicate increasing export concentration in recent years. This suggests that structural upgrading may be occurring alongside a narrowing export base, implying that diversification into complex products has not yet become broad-based. Uzbekistan and Kazakhstan present a more balanced picture, with modest improvements in sophistication accompanied by relatively stable or moderate concentration levels, which is more consistent with gradual structural diversification.

Georgia and Kyrgyzstan display more incremental dynamics in the ECI analysis. Their export structures have become somewhat more sophisticated, but diversification largely occurs within sectors that remain relatively close to their existing productive structures and only moderately more complex than current exports. This pattern aligns with the earlier results showing relatively stable concentration indicators and only modest increases in export sophistication, pointing to gradual capability accumulation rather than rapid structural upgrading.

In contrast, Azerbaijan and Tajikistan remain more constrained by relatively low levels of economic complexity. In these economies, the distribution of feasible products remains concentrated in lower-complexity segments of the product space, and many technologically more sophisticated activities remain distant from their current capability base. This result is partly consistent with the earlier findings from concentration indicators: Tajikistan’s export structure remains relatively stable but limited in diversification, while Azerbaijan’s export structure continues to be influenced by resource-based specialization. As a result, the set of feasible diversification opportunities remains narrower and concentrated in sectors with relatively limited technological sophistication.

Overall, the ECI analysis complements the empirical results obtained from HHI, Theil, and EXPY indicators. While some countries in the region demonstrate signs of capability accumulation and gradual upgrading, the results suggest that structural transformation remains uneven across the region. In several cases, improvements in export sophistication occur alongside persistent concentration in a limited number of products, indicating that diversification into more complex sectors has not yet translated into broad-based structural change.

Conclusion

This brief examined the evolution of trade diversification, export sophistication, and structural transformation in the South Caucasus and Central Asia between 2019 and 2024. The results show substantial cross-country differences. While some economies maintain relatively diversified export structures, others remain more dependent on a narrow set of products. Export sophistication has improved modestly in several countries, but in some cases, this has coincided with rising export concentration. This does not necessarily indicate a negative development: such a pattern may reflect successful specialization based on comparative advantage or upgrading into higher-value activities. However, when sophistication gains are concentrated in a small number of products or markets, the resulting export structure may remain vulnerable to external shocks and less supportive of broad-based structural transformation.

The analysis also points to uneven progress in productive capabilities across the region. Some countries are gradually expanding the range of products they can competitively produce, while others remain constrained by narrower capability bases.

These results highlight the nuanced relationship between diversification, sophistication, and economic complexity. Diversifying into more complex sectors can strengthen economic resilience by broadening the range of activities an economy can rely on, reducing dependence on a limited set of simple or commodity-based exports, and enhancing the capacity to adapt to changes in demand, prices, or trade routes. In this context, the key policy challenge is not diversification for its own sake, but fostering the development of productive capabilities that enable more sophisticated, adaptable, and resilient export structures over time.

References

• Balland, P.-A., Boschma, R., Crespo, J., & Rigby, D. (2019). Smart specialization policy in the European Union: relatedness, knowledge complexity and regional diversification. Regional Studies, 53(9), 1252–1268.
• Brummitt, C. D., Gómez-Lievano, A., Hausmann, R., & Bonds, M. H. (2018). Machine-learned patterns suggest that diversification drives economic development.
• Hausmann, R., Hwang, J., & Rodrik, D. (2007). What you export matters. Journal of Economic Growth, 12(1), 1–25.
• Hausmann, R., Hidalgo, C., Bustos, S., Coscia, M., Chung, S., Jimenez, J., Simoes, A., & Yıldırım, M. (2014). The Atlas of Economic Complexity: Mapping Paths to Prosperity. MIT Press.
• Hidalgo, C. A., & Hausmann, R. (2009). The building blocks of economic complexity. Proceedings of the National Academy of Sciences, 106(26), 10570–10575.
• International Monetary Fund. (various years). Regional Economic Outlook: Middle East and Central Asia.
• Neffke, F., Hartog, M., Boschma, R., & Henning, M. (2018). Agents of structural change: The role of firms and entrepreneurs in regional diversification.
• Rodrik, D. (2006). What’s so special about China’s exports? China & World Economy, 14(5), 1–19.
• World Bank. (2020). Central Asia Trade: Structural Transformation and Diversification.
• Harvard Growth Lab. (2024). Atlas of Economic Complexity – Feasibility Analysis.
• United Nations Comtrade Database. (2024). International Trade Statistics.
• World Bank. (2023). World Development Indicators.

Disclaimer: Opinions expressed in policy briefs and other publications are those of the authors; they do not necessarily reflect those of the FREE Network and its research institutes.