Domestic Intersectoral Linkages in Ukraine’s Economy: Typology and Structural Gaps

1. Introduction

Ukraine’s economy in 2022 was observed under exceptional conditions, including a severe output decline, disruptions to logistics and shifts in the structure of demand. These circumstances make the most recent Organisation for Economic Co-operation and Development (OECD) input–output table an important diagnostic benchmark and require careful comparison with Ukraine’s longer-term structural trajectory. In this article, 2022 is treated as the latest available cross-sectional representation of Ukraine’s intersectoral configuration. This configuration is interpreted against two earlier Ukrainian benchmarks, 2002 and 2012, and against three European reference economies.

The methodological toolkit for analysing intersectoral linkages has developed within the classical tradition of input–output modelling initiated by Leontief (1986); Chenery and Watanabe (1958) and Rasmussen (1956) and consolidated in the contemporary synthesis of Miller and Blair (2022). The existing literature already provides several relevant lenses: classical linkage indices, weighting and normalisation procedures and the distinction between domestic and total effects. The research gap addressed here is therefore not the absence of these methods as such, but the limited joint application of these lenses to Ukraine’s economy using harmonised OECD data and a comparative benchmark design.

This article constructs a multi-dimensional typology of sectors in Ukraine’s economy by applying classical, economic-weight and extended scale-adjusted versions of the Rasmussen approach to domestic and total Leontief inverse matrices. The study addresses three research questions. RQ1 asks how Ukraine’s 2022 sector typology compares with selected within-country benchmarks for 2002 and 2012. RQ2 asks how the identification of key sectors, sectors with high backward linkages, sectors with high forward linkages and sectors with weak linkages changes when linkage intensity is complemented by sectoral economic weight. RQ3 asks which sectors appear stronger in the total matrix than in the domestic matrix and whether such changes indicate import-mediated upgrading rather than domestic linkage strengthening.

The article contributes to the literature on three levels. Empirically, it provides a current-state typology of Ukraine’s domestic and total intersectoral linkages and compares the latest available Ukrainian configuration with two within-country benchmark years. Comparatively, it places the Ukrainian configuration alongside Germany, Poland and Slovakia, which represent different forms of production-network organisation in Europe. Analytically, it shows that the apparent strengthening of some sectors in the total matrix may reflect the inclusion of imported intermediate inputs rather than a deeper domestic production network. The remainder of the article is organised as follows. Section 2 reviews the relevant literature. Section 3 explains the data and methodological protocol. Section 4 presents the results for Ukraine and the benchmark economies. Section 5 discusses the implications of the findings and Section 6 concludes.

2. Literature Review

The study of intersectoral interactions reached its conceptual maturity with the formulation of the input–output model, which first made it possible to quantitatively describe the system of production dependencies among industries in a national economy (Leontief, 1986). Building on this foundation, Chenery and Watanabe (1958) proposed a classification of sectors by the intensity of their direct forward and backward linkages based on the matrix of technical coefficients, while Rasmussen (1956) shifted the analytical focus from direct dependencies to full intersectoral effects captured through the Leontief inverse matrix. The concept of key sectors developed by Hirschman (1958) endowed these indices with a clear analytical meaning: key sectors are not simply industries with high interaction coefficients but the nodes through which multiplier impulses propagate across the entire economy. In its contemporary form, this tradition is synthesised in Miller and Blair (2022), which remains the methodological standard for the analysis of intersectoral, interregional and international production networks.

However, subsequent developments in the literature have demonstrated that classical intersectoral indices are not neutral with respect to an economy’s structural features. Their key limitation lies in their invariance to the absolute scale of sectors: industries with high technological intensity of intermediate consumption may acquire inflated analytical status even where their contribution to gross value added or final demand remains modest. It is precisely for this reason that the methodological discussion from Hazari (1970) and Cella (1984) through Clements (1990); Cuello, Mansouri and Hewings (1992); Dietzenbacher (1992) and Cai and Leung (2004) has concentrated on ways of correcting the indices through weighting, normalisation and refined criteria for identifying key sectors. These studies have convincingly demonstrated that, without accounting for the economic weight of industries, the classical toolkit tends to reproduce the technological profile of the system rather than its actual macroeconomic hierarchy. Contemporary analysis of intersectoral linkages has therefore moved progressively from a simple measurement of linkage intensity towards a combination of intensity with the economic significance of sectors.

A more specific precedent for the weighting logic used in this article is provided by the Laumas tradition of weighted Rasmussen linkages. In this approach, linkage measures are corrected by the sectoral shares of value added and final demand; Ciaschini et al. (2021), following Laumas (1976), explicitly define backward linkages weighted by value added and forward linkages weighted by final demand. This literature provides the methodological basis for treating economic scale not as a substitute for linkage intensity, but as a complementary dimension of sectoral importance.

A new dimension of this problem emerged under the influence of the globalisation of production and the spread of international supply chains. Studies of vertical specialisation have shown that the rising share of imported intermediates alters not only the foreign-trade profile of countries but the very architecture of their internal production dependencies (Hummels, Ishii & Yi, 2001). Subsequent work by Johnson and Noguera (2012); Koopman, Wang and Wei (2014) and Wang et al. (2017) has persuasively justified the need to distinguish domestic and total effects within the input–output framework, since part of the declared intersectoral interaction may not be realised within the national economy but rather through the imported component of intermediate consumption. For the analysis of structural resilience, this distinction is fundamental: a sector that appears structurally central in the total matrix does not necessarily generate a commensurate internal multiplier effect. Accordingly, the contemporary literature increasingly combines intersectoral analysis with approaches for assessing countries’ participation in global value chains and production fragmentation (Los, Timmer & de Vries, 2015; Miroudot & Nordström, 2020), the localisation of production nodes and the asymmetries of dependence among different links of the system of economic reproduction.

Over the past decade, the Rasmussen method has undergone a new wave of empirical reassessment, in which attention has focused on the qualitative reconsideration of the composition of key sectors. Empirical findings for countries at different levels of development indicate that the move from classical to weighted indices frequently shifts the analytical centre away from traditional material-intensive branches towards infrastructural, trading and service activities, through which a substantial share of gross value added and final demand in fact passes (Ojaleye & Narayanan, 2022; Ramírez-Álvarez et al., 2024). In transitional economies, such changes are interpreted not merely as a statistical weighting artefact but as a manifestation of deeper shifts in the structure of economic reproduction, including premature tertiarisation, the weakening of industrial nodes and the asymmetry between high backward and weak forward linkages (Ahmad & Alvi, 2024; Ibourk & Elouaourti, 2025). In this context, the weighted modifications of the Rasmussen method should be regarded not as a technical supplement to the classical toolkit but as a necessary condition for the correct interpretation of the intersectoral structure of a contemporary economy.

Comparative studies of Central and Eastern Europe complement this discussion with an important spatial dimension. They show that integration into regional and global production chains transforms national intersectoral systems in different ways, depending on the size of the economy, the depth of the domestic industrial network and the role of cross-border fragmentation in the formation of output (Ambroziak, 2018; Hagemejer & Mućk, 2019; Cutrini & Mendez, 2023). Small open economies typically exhibit a higher sensitivity of internal structure to the import component, whereas economies with a more substantial domestic industrial base retain greater stability even under high participation in international production networks. In parallel, network-based approaches grounded in intersectoral matrices demonstrate that the production structure is not a neutral background for macroeconomic dynamics: the centrality of individual nodes shapes the propagation of shocks, productivity impulses and employment effects (Acemoglu et al., 2012; Carvalho, 2014; Fadinger, Ghiglino & Teteryatnikova, 2022; Tsoulfidis & Athanasiadis, 2022; Kazekami, 2024). Intersectoral analysis and network concepts thus do not contradict but mutually reinforce one another.

In the case of Ukraine’s economy, the analysis of intersectoral linkages is complicated by prolonged deindustrialisation, dependence on imported intermediate inputs and the uneven development of domestic production chains. The 2022 table adds a current diagnostic point observed under extraordinary conditions, but it should be read together with earlier structural benchmarks. Despite existing studies on particular aspects of Ukraine’s structural development, the literature still lacks a systematic application of classical, economic-weight and extended Rasmussen-type indicators to both domestic and total Leontief matrices for Ukraine in a within-country and European comparative perspective. The present study is aimed at filling this empirical and analytical gap.

3. Methodology

The empirical basis of the study consists of national input–output tables drawn from the OECD Input–Output database. For Ukraine, the analysis uses domestic and total Leontief inverse matrices for 2002, 2012 and 2022 constructed from the harmonised national tables included in the 2025 edition of the database. These data distinguish 50 economic activities and report domestic and imported intermediate flows separately. The results for Ukraine in 2022 are compared with the corresponding indicators for 2002 and 2012 to identify persistent features and recent changes in the country’s intersectoral structure. In addition, the results obtained using the classical method are compared with the corresponding indicators for Germany, Poland and Slovakia, calculated from their 2020 national input–output tables included in the 2023 edition of the OECD database. The corresponding 2022 domestic and total calculations are reported in Appendices A1 and A2. To ensure comparability, the sector classifications are harmonised according to the aggregated activity categories reported in Appendix A3. Germany represents an economy with a deep domestic industrial network, Poland an economy with active participation in regional production chains, and Slovakia a small open economy with a high degree of cross-border integration. This configuration establishes a comparative context in which the results for Ukraine can be interpreted not in isolation but within a broader system of structural differences.

The methodological framework comprises three sequentially complementary modifications of the Rasmussen method: the classical, the economic-weight and the extended scale-adjusted dimensions. The classical modification (Rasmussen, 1956; Miller & Blair, 2022) is based on the Leontief inverse matrix, whose coefficients reflect the full requirement for output of sector i per unit of final demand of sector j. For total linkages, the matrix is defined as:

For domestic linkages, the analogous Leontief inverse is obtained by replacing A with the matrix of domestic technical coefficients A^d:

In Eqs (1) and (2), I denotes the identity matrix of dimension n × n (with n = 50 sectors); A is the matrix of total intermediate input coefficients, whose elements a_ij capture the inputs of sector i per unit of output of sector j incorporating both domestic and imported supplies; A_d is the matrix of domestic intermediate input coefficients with external flows removed; and l_ij are the elements of the Leontief inverse, reflecting the total direct and indirect requirement for output of sector i needed to satisfy one unit of final demand of sector j.

On the basis of the Leontief inverse, normalised backward (BL) and forward (FL) linkage indices are computed as:

In Eqs (3) and (4) BL_j^R denotes the normalised backward linkage index of sector j, capturing the average intensity of its demand for intermediate inputs from the rest of the economy per unit of its own output, while FL_i^R denotes the normalised forward linkage index of sector i, capturing the average intensity of its supplies to other sectors. Normalisation is performed relative to the system-wide mean of the elements of the Leontief inverse. An index value above unity indicates that the sector has an above-average level of full multiplier influence on the economic system. On this basis, all sectors are allocated to four classification groups: key sectors (K) exceeding the threshold in both dimensions; sectors with high backward linkages (B); sectors with high forward linkages (F); and sectors with weak linkages (W).

The economic-weight dimension captures the macroeconomic scale of sectors. The backward-side weight is defined as the normalised sector share in total gross value added, while the forward-side weight is defined as the normalised sector share in final consumption. Both weights are scaled to have a mean value of unity across sectors. For the domestic matrix, final consumption includes domestic household final consumption expenditure, final consumption of non-profit institutions serving households and general government final consumption. For the total matrix, the corresponding domestic-plus-imported final-consumption component is used. Gross fixed capital formation, exports and inventory changes are excluded in both cases:

The choice of weights follows the weighted-linkage tradition while adapting it to the domestic-demand focus of the present study. Laumas (1976) and Ciaschini et al. (2021) show that Rasmussen linkages can be adjusted by sectoral shares of value added and final demand; in particular, Ciaschini et al. (2021) define backward linkages weighted by value added and forward linkages weighted by final demand. In the present article, the share of a sector in gross value added captures its realised contribution to domestic income creation, while the share in final consumption captures its role in meeting final demand. The domestic and total specifications of final consumption (FC_i) allow the analysis to distinguish domestic-demand significance from domestic-plus-imported final-consumption significance.

The extended scale-adjusted indicators combine the classical linkage intensity with the corresponding economic-weight dimension. They are computed as normalised products of the classical Rasmussen index and the relevant weight:

In Eqs (7) and (8) EBL_j and EFL_i are the extended normalised backward and forward linkage indices, respectively. Their mean value equals unity, which makes the threshold of 1.0 directly comparable with the classical Rasmussen convention. The typology of sectors in the extended dimension therefore follows the same four-group rule as the classical method. To quantify the impact of import dependence, pairwise differences are computed for each sector between values obtained from the domestic and total Leontief inverse matrices, while the qualitative transformation is formalised as a change in the sector classification position. For cross-country benchmarking and within-Ukraine comparison, the shares of sectors that change their classification position are computed along with the qualitative direction of these transitions.

4. Results

The results are presented in four steps. First, the baseline 2022 configuration of full multiplier linkages is reported for the classical, economic-weight and extended dimensions. Second, the domestic and total matrices are compared in order to identify sectors whose position changes when imported intermediate inputs are included. Third, Ukraine’s 2022 profile is benchmarked against Germany, Poland and Slovakia. Fourth, a within-Ukraine comparison for 2002, 2012 and 2022 is added to distinguish persistent features of Ukraine’s intersectoral structure from sector-role changes that are more specific to the latest available benchmark. The full numerical values for Ukraine in 2022 are reported in Appendices A1 and A2, the cross-country classification matrix in Appendix A3 and the additional Ukrainian benchmark calculations for 2002 and 2012 in Appendices A4–A7. The baseline 2022 typology is presented in Table 1.

According to the results obtained using the classical method, the leading positions in Ukraine’s intersectoral interaction are concentrated in resource extraction and primary processing, represented most clearly by Crop and animal production, hunting and related service activities (A01), Extraction of crude petroleum and natural gas (B06) and Basic iron and steel (C24A); the full composition of the group is reported in Table 1. Equally important is the fact that sectors with high backward linkages form the largest group and are concentrated in manufacturing. A substantial part of Ukraine’s industrial segment therefore functions primarily as a system of intensive intermediate consumption, while its capacity to generate stable forward effects for other industries is comparatively weaker. This configuration points to a structural imbalance between the phase of resource mobilisation and the propagation of production outcomes through domestic chains.

The results obtained in the economic-weight dimension differ substantially from those obtained using the classical method and indicate a different centre of gravity of the economy. Public administration and defence (O), Education (P) and Human health and social work activities (Q), which were not part of the leading group according to the classical method, acquire the status of key sectors. This result reflects the high shares of these sectors in gross value added and domestic final consumption, which become central once the analysis moves from pure linkage intensity to scale-adjusted structural significance. Alongside them, the scale-adjusted core includes agriculture and food production, energy, trade and transport, financial services and real estate activities (Table 1).

The differences among the classifications obtained using the classical method and in the economic-weight and extended dimensions are substantial. When moving from the classical to the weighted approach, 35 of 50 sectors (70%) change their classification position, compared with 34 of 50 sectors (68%) when moving from the classical to the extended approach. By contrast, the weighted and extended modifications produce almost identical outcomes. The decisive factor reshaping the intersectoral structure is therefore the introduction of the economic weight of sectors rather than any further technical refinement of the synthetic index. The transitions from the group of sectors with weak linkages to key sector status for Public administration and defence; compulsory social security (O), Education (P), Human health and social work activities (Q), Financial and insurance activities (K) and Real estate activities (L) are particularly telling: all of them indicate the systematic underestimation by the classical approach of activities with relatively low technological intensity but high macroeconomic centrality. The reverse transitions from key to weak are concentrated in parts of the resource-extraction and primary-processing industries, suggesting that their importance is overstated in the unweighted measurement. The group of sectors with high forward linkages deserves particular attention. According to the classical method, it has a clearly infrastructural–service profile and comprises trade, transport activities, energy supply, finance and professional and information services. In the weighted and extended approaches, however, it virtually disappears as a separate category, with most of these sectors moving into the group of key sectors. In the actual structure of Ukraine’s economy a substantial share of ‘forward’ sectors are therefore not one-dimensional suppliers to other industries but simultaneously accumulate significant backward effects through their own scale in domestic demand and value creation. Moreover, the sharp rise in the share of sectors with weak linkages in the weighted and extended dimensions reflects not analytical redundancy of these approaches but a high concentration of multiplier potential within a relatively narrow set of activities. The Ukrainian economy is thus characterised not by an even distribution of structural significance but by a pronounced hierarchy of sectoral roles.

The second analytical direction concerns the impact of import dependence, that is, the changes that emerge in the move from the domestic to the total Leontief inverse matrix. According to the classical method, only 10 sectors, or 20% of the sample, change their classification position. This provides grounds for speaking of a relative inertia of full intersectoral linkages in Ukraine’s economy: even the inclusion of the import component does not produce a mass shift in sectoral functional roles. The interpretation of this inertia is not univocal, however: it may simultaneously reflect the internal cohesion of core segments and a limited depth of integration into global production networks. The directions of these changes are summarised in Table 2.

The qualitative analysis of the changes between the domestic and total matrices, summarised in Table 2, reveals two opposing tendencies. The first tendency is associated with the weakening of classification positions in sectors such as Forestry and logging (A02), Accommodation and food service activities (I), IT and other information services (J62–J63) and Financial and insurance activities (K). In these sectors, the inclusion of the import component does not strengthen but rather reduces internal intersectoral centrality, indicating either a partial displacement of domestic linkages by external resources or a sectoral orientation beyond the internal production contour. The second tendency is conceptually more significant: several sectors acquire higher classification positions in the total matrix, with the most informative cases being Chemicals and chemical products (C20), Computer, electronic and optical products (C26) and, especially, Water transport (H50); the full set of transitions is reported in Table 2. Yet these changes should not be read as an automatic strengthening of domestic production linkages. Rather, they reflect an import-mediated expansion of the intersectoral role, whereby a sector becomes more important in the total system not because it becomes more deeply integrated into national reproduction but because it functions as a channel for the passage of external resources. This is the phenomenon interpreted in the present article as the effect of ‘import-mediated upgrading’.

The comparison with Germany, Poland and Slovakia, detailed in Table 3, allows for a more precise interpretation of Ukraine’s structural profile. The share of sectors changing their classification position in the transition from the domestic to the total matrix amounts to 20% in Ukraine, 29% in Poland, 31% in Germany and 42% in Slovakia. Ukraine’s economy thus exhibits the lowest share of sectors changing classification among the comparator countries. This result should be interpreted as a descriptive feature of the 2022 benchmark rather than as evidence of a causal wartime transformation. Germany’s relative stability is consistent with a deep and diversified domestic industrial network in which imported inputs complement rather than replace domestic linkages. Slovakia’s high sensitivity can be interpreted in light of its assembly-type integration into cross-border production chains, where the functional role of many manufacturing activities emerges only once imported inputs are taken into account. Poland occupies an intermediate position, combining a sizeable domestic industrial base with active participation in regional production networks. Ukraine’s low reclassification share should therefore not be read as a sign of structural strength: it is consistent with a narrower domestic production contour and more selective import embeddedness, in which fewer sectors are integrated into external supply systems deeply enough for the import component to change their functional role. The sector-level alignment in Table 3 and Appendix A3 makes this contrast concrete. In Germany and Poland, Land transport and transport via pipelines (H49) and Warehousing and support activities for transportation (H52) hold key-sector status in both the domestic and total matrices, whereas in Ukraine they remain in the group of sectors with high forward linkages. Conversely, the extractive and fuel-processing activities identified as key sectors for Ukraine’s domestic matrix using the classical method do not attain key-sector status in the domestic matrix of any comparator country; in Poland and Slovakia, Coke and refined petroleum products (C19) become a key sector only in the total matrix, that is, through imported inputs. In Slovakia, the transition to the total matrix is associated with higher classification positions for a broad group of manufacturing activities, while in Ukraine comparable changes are much narrower and coexist with lower positions for information and financial services.

The detailed alignment of classification positions in Table 3 allows for three dimensions of Ukraine’s specificity to be distinguished. The first concerns sectors with a classification position unique to Ukraine under the domestic matrix. This applies above all to six industries of the resource-extraction complex, which acquire key sector status only in Ukraine’s economy, thereby quantitatively confirming the resource-based composition of the group of key sectors identified using the classical method. Yet the logistical segment, represented by Land transport and transport via pipelines (H49) and Warehousing and support activities for transportation (H52), does not attain in Ukraine the full key sector status observed in the comparator countries, indicating an insufficient depth of its two-sided integration into production chains. Finally, Water transport (H50), Air transport (H51), Real estate activities (L) and Water supply; sewerage, waste management and remediation activities (E) are classified as sectors with weak linkages, which points to the peripheral position of part of the infrastructural complex within the internal architecture of intersectoral linkages.

The second dimension of specificity concerns sectors for which the change in classification position during the transition to the total matrix is characteristic only of Ukraine. Water transport (H50) moves from the group of sectors with weak linkages to key sector status, and Air transport (H51) to the group of sectors with high backward linkages, reflecting their role as channels associated with external flows. Forestry and logging (A02) and IT and other information services (J62–J63), by contrast, show a decline in status, indicating the vulnerability of their internal linkages to changes in the external contour. The third dimension covers sectors that preserve classification stability in Ukraine while two or three comparator countries exhibit transformations. This invariance is concentrated in the raw-material-processing segment and partly in the production-technological segment, pointing to the relative rigidity of certain structural nodes. Taken together, these dimensions show that Ukraine’s model does not replicate any of the comparator-country models but represents a distinct current configuration of intersectoral organisation.

The synthesis of the 2022 results reveals four systemic features of Ukraine’s intersectoral structure. First, multiplier potential is concentrated within a narrow set of sectors, although different centres of this concentration are identified using the classical and scale-adjusted approaches. Second, manufacturing tends towards high backward and comparatively weaker forward effects, which limits the depth of internal propagation of production impulses. Third, parts of the infrastructural and service segments remain insufficiently embedded in the internal production contour, although some of these activities become important once economic weight is taken into account. Fourth, some upgrades in the total matrix are better interpreted as import-mediated upgrading than as domestic linkage strengthening. These findings justify the additional within-Ukraine benchmark comparison reported in Table 4.

The benchmark comparison indicates that the sharp difference between classical and scale-adjusted classifications is a persistent feature of Ukraine’s intersectoral structure rather than a result that appears only in 2022. In 2002 and 2012, as in 2022, key-sector status is concentrated in a narrower set of activities in the economic-weight and extended dimensions than under the classical method. At the same time, the domestic-total gap became more visible in 2022. According to the classical method, the number of sectors changing classification between the domestic and total matrices increased from 8 of 50 sectors in 2002 and 7 of 50 sectors in 2012 to 10 of 50 sectors in 2022. In the economic-weight dimension, changes were limited to one sector in 2002 and one sector in 2012 and were absent in 2022. In the extended dimension, the corresponding number rose from one sector in 2002 and three sectors in 2012 to four sectors in 2022. This means that the import component affected the qualitative identification of sectoral roles more strongly in 2022 than in the two earlier benchmarks, although its effect remained smaller than the reclassification caused by introducing economic scale into the analysis.

According to the results obtained from the domestic matrix using the classical method, only Crop and animal production, hunting and related service activities (A01), Other non-metallic mineral products (C23) and Basic iron and steel (C24A) retain key sector status throughout 2002, 2012 and 2022. By 2022, however, several extractive and primary-processing activities had entered or returned to the group of key sectors, most notably Mining of coal and lignite (B05) and Coke and refined petroleum products (C19); the full composition of the group is reported in Appendices A1, A4 and A6. At the same time, several manufacturing branches occupy weaker positions than in earlier years: most notably, Food products, beverages and tobacco (C10–C12) and Chemicals and chemical products (C20) move from key sector status to the group of sectors with high backward linkages, while Pharmaceuticals, medicinal chemical and botanical products (C21) and Computer, electronic and optical products (C26) fall to the group of sectors with weak linkages. This means that the 2022 configuration obtained from the domestic matrix using the classical method is more concentrated around resource extraction and basic processing, whereas parts of manufacturing display weaker two-sided domestic linkage positions than in the earlier benchmarks.

A different pattern is observed in the economic-weight and extended dimensions. The group of key sectors identified in these dimensions is more stable over time than the group identified using the classical method. In the domestic economic-weight dimension, the same 10 activities spanning agriculture and food production (A01, C10–C12), energy supply (D), trade and land transport (G, H49), finance and real estate (K, L), and public administration, education and health care (O, P, Q) remain key sectors across all three benchmark years. In the extended dimension, almost the same group is identified using the domestic matrix, with Electricity, gas, steam and air conditioning supply (D) returning to key sector status in 2022 after a weaker position in 2012. This indicates that the 2022 prominence of these activities is not merely a 1-year classification artefact, but reflects their high macroeconomic weight within Ukraine’s intersectoral structure.

Comparison of the results obtained from the domestic and total matrices provides a further distinction by showing how sector roles change once imported intermediate inputs are included in the full requirements system. In 2002, higher classification positions were recorded mainly for extractive, fuel-related and transport activities, whereas lower positions were observed primarily for Forestry and logging (A02) and Professional, scientific and technical activities (M). In 2012, the changes were fewer and more mixed, combining stronger positions of fuel-related activities with lower positions in several service activities. By contrast, in 2022, import-mediated upgrading became more visible in manufacturing and transport-related activities, most notably Chemicals and chemical products (C20), Computer, electronic and optical products (C26) and Water transport (H50); the full lists of transitions can be reconstructed from Appendices A2, A5 and A7. This confirms that several sectors have stronger apparent positions in the 2022 total matrix, despite weaker domestic linkage positions.

The sectors most adversely affected by the import-mediated comparison are those whose classification position falls when the analysis moves from the domestic to the total matrix. In 2022, according to the classical method, this group includes Forestry and logging (A02), Accommodation and food service activities (I), IT and other information services (J62–J63), and Financial and insurance activities (K). In the extended dimension, it includes Electricity, gas, steam and air conditioning supply (D), Education (P) and Human health and social work activities (Q). All these sectors move to a weaker classification category (Table 2). These changes indicate that once imported flows are included, the relative domestic centrality of these sectors is diluted. Compared with 2002 and 2012, the 2022 pattern spans a broader range of activities, including energy supply, information services, financial activities, education and health care.

5. Discussion

The results provide a diagnostic assessment of Ukraine’s intersectoral configuration. The within-Ukraine benchmarks show that several features visible in the latest cross-section—especially the divergence between classical and scale-adjusted classifications and the concentration of key roles in a limited group of sectors—were already present in earlier benchmark years. The sector-level comparison also shows that the composition of the group of key sectors identified for 2022 using the classical method changed: persistent key sectors such as Crop and animal production, hunting and related service activities (A01), Other non-metallic mineral products (C23) and Basic iron and steel (C24A) are joined by a stronger extractive and primary-processing group, while several manufacturing branches lose two-sided domestic linkage status. By contrast, the results obtained in the economic-weight and extended dimensions indicate a more stable macroeconomically significant group comprising agriculture and food production, energy, trade and transport, financial services, real estate activities, public administration, education and health care (Table 1). This distinction separates persistent structural features from the specific sector-role configuration captured by the latest OECD table. The comparison with Germany, Poland and Slovakia shows that Ukraine has a lower share of sectors changing classification between the domestic and total matrices. Combined with the key-sector status of several extractive and primary-processing activities, this is consistent with a stronger concentration of intersectoral linkages in a relatively narrow resource-based production core.

The strongest empirical result is the difference between pure linkage intensity and economic scale. In 2022, 70% of sectors change classification when the analysis moves from the classical to the economic-weight dimension. This does not mean that the classical Rasmussen method is inadequate in general; rather, it shows that different indicators answer different questions. Classical indices identify technologically intensive intersectoral positions, whereas sectors combining linkage roles with macroeconomic scale occupy more prominent positions in the economic-weight and extended dimensions. For Ukraine, this distinction is crucial because resource-extraction and primary-processing activities dominate the group of key sectors identified using the classical method, while agriculture and food production, energy, trade and transport, financial services, real estate activities, public administration, education and health care become prominent after scale adjustment.

The domestic-total comparison provides a second important finding. Some sectors improve their classification position in the total matrix, although this does not necessarily indicate a strengthening of domestic production linkages. The within-Ukraine comparison shows that this import-mediated effect intensified in 2022 relative to the two earlier benchmarks: the share of sectors changing classification between the domestic and total matrices according to the classical method increased from 16% in 2002 and 14% in 2012 to 20% in 2022. Classification changes in the economic-weight dimension were marginal in 2002 and 2012 and absent in 2022, while the share of changes in the extended dimension rose from 2% in 2002 and 6% in 2012 to 8% in 2022. The composition of these changes also shifted. In 2002 and 2012, import-mediated upgrading was concentrated mainly in extractive, fuel-related and selected transport or metals activities. In 2022, it became more visible in chemicals and chemical products, pharmaceuticals, rubber and plastics products, computer, electronic and optical products and transport activities. Conversely, forestry and logging, accommodation and food services, information services, financial activities, energy supply, education and health care show downgrades in at least one analytical dimension (Table 2). The term ‘import-mediated upgrading’ is used for cases in which a sector’s position improves in the total matrix without an equivalent strengthening in the domestic matrix. For reconstruction policy, this distinction implies that sectors should not be prioritised solely because they occupy a strong position in the total matrix; priority should be given to activities capable of converting import-mediated linkages into domestic intermediate supply and stronger internal multiplier effects.

Taken together, the results support the joint use of classical, economic-weight and extended indicators for diagnosing sectoral roles. The contribution of the article lies in applying linkage, weighting and domestic-total decomposition tools jointly to Ukraine and in showing how the interpretation of sectoral roles changes across these analytical lenses. This provides an empirical basis for structural diagnosis and for future work on the dynamics of Ukraine’s production network.

6. Conclusions

The study provides four main conclusions on the intersectoral structure of Ukraine’s economy. First, different groups of structurally important sectors are identified using the classical and scale-adjusted approaches. According to the classical method, the leading positions are occupied mainly by resource extraction and primary-processing activities, whereas in the economic-weight and extended dimensions, the main structurally significant activities comprise agriculture and food production, energy, trade and land transport, financial and real estate activities, public administration, education and health care. The within-Ukraine comparison further shows that, by 2022, the group of key sectors identified from the domestic matrix using the classical method became more concentrated around resource extraction and primary processing, while several manufacturing branches occupied weaker two-sided positions than in 2002 and 2012. Second, the group of key sectors identified in the economic-weight dimension is more stable over time: agriculture and food production, energy supply, trade and land transport, financial and real estate activities, public administration, education and health care remain central across all three benchmark years. Third, the import-mediated component became more visible in 2022 than in the earlier Ukrainian benchmarks. The share of sectors changing classification between the domestic and total matrices according to the classical method reached its highest value in 2022. The same tendency was observed in the extended dimension on a smaller scale, while the economic-weight dimension remained essentially unchanged. Fourth, the sectors most exposed to import-mediated weakening in 2022 include information services, financial activities, energy supply, education, health care, forestry and logging and accommodation and food services, while import-mediated upgrading is concentrated in chemicals and chemical products, pharmaceuticals, rubber and plastics products, computer, electronic and optical products, and water and air transport. The 2022 configuration therefore combines persistent structural features with a stronger import-mediated layer.

The novelty of the study is primarily empirical and diagnostic. It jointly applies classical, economic-weight and extended Rasmussen-type indicators to domestic and total Leontief inverse matrices for Ukraine, adds within-country benchmark years for 2002 and 2012 and places the latest Ukrainian configuration in a European comparative context. The analysis maps the current configuration and shows which sector roles were stable over time, which became more visible in 2022 and which appear stronger only in the total matrix.

The policy relevance of the results lies in providing a diagnostic basis for structural and reconstruction policy. Sectors that combine substantial domestic linkages with economic scale deserve particular attention because they may transmit demand and investment impulses more strongly through the national economy. Sectors displaying import-mediated upgrading require a different interpretation: their apparent strength in the total matrix may point to opportunities for developing domestic intermediate supply, but it should not be treated as evidence of already strong domestic production chains. This is particularly relevant for chemicals and chemical products, pharmaceuticals, rubber and plastics products, computer, electronic and optical products, and water and air transport. Conversely, sectors whose position deteriorates when moving from the domestic to the total matrix—especially information services, financial activities, accommodation and food services, forestry and logging, energy supply, education and health care—require a more detailed assessment of whether imported inputs substitute for, complement or weaken domestic intersectoral linkages. The study is subject to three principal limitations. First, the use of selected benchmark years does not replace a full annual dynamic analysis. Second, the level of sectoral disaggregation to 50 activities smooths out internal heterogeneity. Third, the threshold value of 1.0 in the typology simplifies the real gradations of sectoral roles. Further research should extend the analysis to a full panel and combine the extended Rasmussen approach with network indicators of centrality and resilience.