A new economic modeling framework that blends unit structure analysis with global production networks has mapped how Japan’s automotive sector relies on complex domestic industrial clusters within foreign trading partners. Developed by researcher Jun Omata and published in the Journal of Economic Structures, the study demonstrates that the intermediate value added supporting Japanese automobile assembly is not merely channeled through isolated exporting facilities abroad, but rather originates within dense, interconnected domestic supplier networks in partner countries. Applying this method to international trade and production data spanning from 2002 to 2022 reveals a profound structural shift in how vehicle manufacturing functions across borders. In particular, the findings illustrate how the rapid expansion of China’s domestic industrial base has integrated directly into the upstream supply pipelines of Japanese vehicle production, while the United States has simultaneously maintained and deepened extensive domestic interindustry linkages that feed components into Japanese manufacturing systems.
Understanding international production networks has become critical as modern manufacturing processes increasingly offshore distinct operational stages to specialized facilities across the globe. Traditional international trade statistics, which rely primarily on gross customs records, routinely track the movement of goods across political boundaries but obscure the true origin of intermediate inputs and the geographic distribution of economic value added. To address this limitation, previous foundational research in global value chain analysis developed mathematical formulations based on the Leontief inverse matrix to convert standard customs transactions into value-added flows. While pioneering frameworks developed by earlier economists succeeded in tracing the total income generated in source economies by allocating final product demand to contributing countries, they conceptualized destination economies strictly as final consumers or end-product producers. Consequently, these established trade-accounting frameworks lacked the structural granularity needed to uncover how upstream intermediate stages connect through local industrial groupings prior to final assembly.
To overcome these empirical and conceptual blind spots, the new investigation extends a classic technological approach originally formulated by Ozaki in 1980, known as unit structure analysis. In standard input-output economics, the classical Leontief inverse matrix calculates how final demand across various sectors propagates through an economy, arranging induced output along rows attributed to the columns of originating final demand. Ozaki’s unit structure system departs from this convention by recording induced intermediate production within the specific columns of the industries and countries that directly generate that output, rather than subsuming it under the final demand column. By organizing intermediate transactions associated with a given commodity into an explicit input-output table format, unit structure analysis makes it possible to isolate the exact interindustry relationships and technical coefficients underlying any single unit of final output. Building upon recent extensions by Kuroiwa, the current study converts these intermediate output transactions into a unified global value network.
The mathematical formulation of this value network model establishes an exact distinction between intermediate production stages and direct final assembly. By multiplying the input coefficient matrix and the Leontief inverse column vector of a designated industry by a diagonalized value-added coefficient matrix, the model generates an adjacency matrix representing the unit value network. When scaled by total final demand, this yields a total value network matrix that maps all value-added flows generated across intermediate production processes that are indirectly induced by that specific final demand. Unlike previous models that define destination countries exclusively as final consumers, this approach explicitly identifies destination countries as importers participating in subsequent manufacturing tiers. As demonstrated through the formal algebraic identity in the study, the row-wise sum of this intermediate value-added network precisely equals the total value added embodied in the final good minus the value added generated during direct final-stage production within the assembling country.
The operational mechanics of the model are illustrated in the study through a multi-country production system involving synthetic resins, automotive components, and completed passenger vehicles. In this baseline conceptual example, Country A manufactures synthetic resins valued at 10 units and exports them to Country B, which consumes the resin to produce tires valued at 40 units. Simultaneously, Country C manufactures camshafts valued at 60 units and pistons valued at 30 units, using both intermediate inputs to construct vehicle engines valued at 160 units. Country B and Country C then export their finished tires and engines to Country D, which performs final automobile assembly valued at 350 units for local sale. Under the extended unit structure framework, Country D’s final demand of 350 units corresponds exactly to 350 units of total value added distributed across the four participating nations, with 200 units generated across intermediate upstream production stages and 150 units generated directly during final vehicle assembly in Country D.
To evaluate these relationships empirically across real-world trading corridors, the research evaluated the motor vehicles, trailers, and semi-trailers sector, cataloged as Code 29, across three benchmark years: 2002, 2012, and 2022. The empirical investigation drew upon the 2025 edition of the OECD Inter-Country Input-Output tables, which encompass 50 economic sectors across 80 countries alongside an aggregate rest-of-the-world region. According to gross output figures calculated from these tables, the output of Japan’s automotive sector exceeded 310 billion US dollars in 2002, climbed to 540 billion US dollars in 2012, and subsequently contracted to 380 billion US dollars by 2022, when measured in current prices. The author selected the OECD dataset over alternative resources, such as the Asian Development Bank tables, because the OECD provides a dedicated classification specifically isolating automobile manufacturing, rather than grouping road vehicles into a broader transport equipment category that incorporates shipbuilding and railway rolling stock.
Visualizing the vast matrix of intermediate dependencies across 81 territorial entities and 50 industrial sectors presented significant technical hurdles. Historical applications of unit structure systems, including recent analyses of European Union economies, have predominantly visualized intersectoral dependencies using three-dimensional bar charts that plot intermediate sectors along the horizontal axes and transaction magnitudes on the vertical axis. However, as the author notes, three-dimensional bar charts and dense heat maps become visually illegible when applied to multiregional tables featuring dozens of countries and thousands of intersecting sectoral cells. To enhance visual clarity and make the underlying structural properties interpretable, the investigation adopted directed network diagrams following methodologies pioneered by Kagawa and colleagues. While directed graphs do not display the full absolute magnitude of every microscopic flow, they treat the unit value network as a weighted adjacency matrix, effectively highlighting the architectural backbone of international industrial dependencies.
The study implemented network thresholds to isolate major supply linkages, revealing how technical analytical cutoffs alter the observed structure of global manufacturing. When the researchers extracted international value-added flows supporting Japanese vehicle assembly in 2012 using a higher threshold of 0.0005, corresponding to transactions exceeding roughly 270 million US dollars in current prices, the resulting graph depicted foreign industries connecting almost exclusively through direct bilateral transactions into Japanese manufacturing nodes. However, when the analytical threshold was relaxed to 0.0001, corresponding to intermediate value-added flows exceeding approximately 31 million US dollars in 2002, 54 million US dollars in 2012, and 38 million US dollars in 2022, a far more intricate structural reality emerged. Lowering the threshold exposed dense, highly integrated domestic interindustry networks operating within partner countries, showing that Japanese automobile assembly draws upon broad national industrial bases rather than isolated export enclaves.
Examining bilateral network graphs between Japan and the United States demonstrates that American domestic industrial networks were already heavily engaged in supplying Japanese automotive production in 2002. Over the two-decade study period, Japan’s dependence on American industrial inputs expanded, with value-added linkages increasingly permeating the domestic supplier groups within the United States. Even as Japan’s gross automotive output declined in current-price terms between 2012 and 2022, and despite Japan’s overall import share from the United States remaining relatively stable in national trade statistics, the underlying value-added data indicate that vertical manufacturing stages continued shifting overseas. Furthermore, incoming American value added flowed not only into final motor vehicle assembly in Japan but also into ancillary domestic Japanese industrial sectors, confirming that intermediate production processes across related manufacturing tiers have undergone progressive offshore integration.
The empirical trajectory between Japan and China reveals an even more dramatic industrial transformation over the same twenty-year timeline. In 2002, intermediate industrial contributions from Chinese sectors toward Japan’s automotive production were minimal in value-added terms. By 2012, and accelerating markedly into 2022, both China’s internal value-added networks and its direct transmissions to Japan underwent massive expansion, culminating in a significant surge across the automotive supply network. These computational findings align with earlier output-based input-output studies while demonstrating that China’s emergence as an essential supplier for Japanese automotive manufacturing stems from the formation of an interconnected domestic manufacturing web. Rather than engaging in global value chains through simple resource extraction or fragmented processing stages, China successfully cultivated comprehensive interindustry networks that collectively channel substantial economic value into complex downstream assembly systems abroad.
While the study provides a refined architectural view of cross-border manufacturing, the author highlights important methodological considerations and empirical limitations. Because the empirical calculations were performed using current-price input-output tables, shifts in transaction values over time reflect fluctuating price levels and exchange-rate variations alongside genuine physical adjustments in intermediate supply linkages. The author emphasizes that caution must be exercised when interpreting decade-scale trends without constant-price adjustments. Additionally, because the analytical visualizations focus primarily on bilateral relationships linking Japan to the United States and China, the study does not capture all multilateral transshipments that pass through third-party intermediary states. All calculations were executed using the open-source statistical programming language R, version 4.5.0, confirming that the analytical framework can be readily adapted to evaluate other domestic industries or extended to investigate global value networks originating in other economies.
Subject of Research: Global value network analysis of Japan's automobile industry using input-output tables
Article Title: Extending unit structure analysis to global value network approach: the case of the Japanese automobile industry
Article References: Omata, J. (2026). Extending unit structure analysis to global value network approach: the case of the Japanese automobile industry. Journal of Economic Structures, 15(1), Article 17. https://doi.org/10.1186/s40008-026-00380-x
Image Credits: AI Generated
DOI: 10.1186/s40008-026-00380-x
Keywords: global value networks, unit structure analysis, Japanese automobile industry, input-output economics, OECD ICIO tables, global supply chains, intermediate transactions, value added flows, offshoring, bilateral trade, industrial networks, Leontief inverse matrix, directed network diagrams, international production networks, intermediate production, trade in value added, gross output
Cite Scienmag News
Reid Dalton. (November 17, 2026). Global Value Network Model Maps Japan’s Offshored Auto Supply Chains. Scienmag. https://scienmag.com/global-value-network-model-maps-japans-offshored-auto-supply-chains/
Reid Dalton. "Global Value Network Model Maps Japan’s Offshored Auto Supply Chains." Scienmag, 17 November 2026, https://scienmag.com/global-value-network-model-maps-japans-offshored-auto-supply-chains/. Accessed 17 November 2026.
Reid Dalton. "Global Value Network Model Maps Japan’s Offshored Auto Supply Chains." Scienmag. November 17, 2026. https://scienmag.com/global-value-network-model-maps-japans-offshored-auto-supply-chains/








