Experimental estimates suggest limited changes in global value chain (GVC) integration in 2023-24 pointing to a slower pace of “deglobalisation” and “reshoring” compared with the period between 2020 and 2022. Domestic value-added share in exports rose marginally from 77% in 2022 to 77.6% in 2024.
Country heterogeneity is substantial: 11 economies increased domestic content by more than 2.5 percentage points from 2022 to 2024, while others remained flat or declined.
Domestic value added in manufacturing stabilised at 62-65% in 2023-24 on average in OECD countries and 75-77% in large emerging economies, after having fluctuated in 2021-22. Services content in manufacturing continued to expand, with most subsectors showing increases in both domestic and foreign services value added.
Small, specialised economies exhibit larger nowcast errors, reflecting structural complexity and idiosyncratic shocks.
Global value chain repositioning
Economy-wide overview
Experimental estimates of five key Trade in Value Added (TiVA) indicators provide little evidence of widespread reshoring in 2023-24. Averaging across 41 economies, export-weighted domestic value-added share in gross exports is estimated to have risen only slightly – from about 77% in 2022 to 77.7% in 2023 and 77.6% in 2024 (Figure 1). Unweighted averages show a similarly modest increase, rising from roughly 71.1% to 72.1% in 2024 over the same period, while foreign value-added shares saw a small decline (from 28.9% to 27.9%). The median domestic value added across 41 economies climbed from about 71.7% in 2022 to 74.2% in 2024. Part of this modest increase in domestic value added likely reflects the rising role of services in production, which are more often sourced locally, rather than systematic reshoring of manufacturing stages back home.
Export-weighted domestic value-added share in gross exports, % of gross exports
Note: The large emerging economies include Brazil, China, India, Indonesia and South Africa.
Source: OECD TiVA 2026 nowcasting exercise.
Among OECD economies, experimental estimates of domestic value-added share show a gradual but steady rise, from an average of 69.7% in 2022 to 70.6% in 2023, reaching 70.8% in 2024, indicating incremental rather than rapid shifts toward domestic sourcing. By contrast, the five large emerging markets covered in this update (Brazil, China, India, Indonesia, and South Africa) display a higher base level of domestic value-added share, rising from 81.3% in 2022 to 82.2% in 2023, before easing back to 81.6% in 2024. These patterns, a small uptick in domestic value added among OECD economies and a largely stable domestic value added share among large emerging economies, appear consistent with OECD analysis of global value chain positions (Andrenelli et al., 2023[1]; Crowe and Rawdanowicz, 2023[2]).
Taken together, these developments suggest limited changes in global value chains (GVC) integration, pointing to a slowing pace of “deglobalisation” and “reshoring” compared with the period 2020-22. Instead, the evidence points to gradual and uneven adjustments in the most recent period. This aligns with recent OECD analysis indicating that firms have focused on reconfiguring supply chains – diversifying suppliers and strengthening inventories – rather than systematically relocating production back home (Schwellnus, Haramboure and Samek, 2023[3]).
Substantial country heterogeneity
Global trends mask wide disparities across countries (Figure 2). A first group of economies – Austria, Belgium, Germany, Hungary, Italy, Latvia, the Netherlands, Portugal, the Slovak Republic, Slovenia and Spain – recorded substantial increases in domestic value-added in gross exports between 2022 and 2024, typically exceeding 2.5 percentage points.
A second group – including Australia, Brazil, Denmark, Indonesia, Ireland, Japan, Norway, South Africa, Sweden, Türkyie, the United Kingdom, and the United States – experienced little to no growth in domestic value-added shares in 2023-24, with changes at or below zero. These flat or declining trends suggest that reliance on foreign inputs has persisted, or even strengthened, in these countries since 2022.
The remaining economies exhibited only moderate increases in domestic value-added shares in 2023-24 – between 0 and 2.5 percentage points – pointing to incremental adjustments rather than a pronounced shift toward greater reliance on domestic sourcing. Their trajectories align with the broader pattern of gradual global value chains reconfiguration, instead of abrupt strategic relocation.
Overall, post-2022 heterogeneity in country adjustment paths is largely reflective of underlying structural and policy differences across countries. Economies with large manufacturing bases – such as Germany or the Slovak Republic, with strong automotive and machinery sectors – and explicit policy efforts to boost resilience or strategic autonomy tend to record larger increases in domestic value added, as firms substitute some foreign inputs with local sourcing and upgrade domestic capabilities (Jaax, Miroudot and Van Lieshout, 2023[4]). Conversely, countries with entrenched comparative advantages in certain tightly integrated global value chains, or where export structures remain heavily oriented toward highly fragmented value chains – like Denmark’s pharmaceutical industry, or Sweden’s machinery and telecommunication equipment chains – generally see flat or declining domestic shares, as established international sourcing patterns persist. Between these extremes, most economies exhibit only modest changes, pointing to gradual, sector-specific fine-tuning of supply chains rather than wholesale reshoring or a broad retreat from globalisation (Andrenelli et al., 2023[1]; OECD, 2025[5]).
In most countries, the services share of domestic value added over the past two years has broadly continued its earlier trend, rising on average by 1.9 percentage points.
Change in domestic value-added share, percentage points
Note: Circles show the change in the domestic value-added share between 2022 and 2023, triangles the change between 2023 and 2024, and squares the cumulative change over 2022–2024.
Source: OECD TiVA 2026 nowcasting exercise.
While aggregate trends remained broadly stable in 2023-24, sectoral differences are evident (Figure 3). In manufacturing, domestic value-added shares appear to have stabilised in recent years after a decline in 2021-22, with OECD economies holding steady in the 62-65% range and large emerging economies in the 75-77% range. Services showed a different pattern: domestic value-added shares fell only marginally in 2022 and have since returned to almost pre-shock levels for both OECD economies and the large emerging markets. This resilience likely reflects the service sector’s relative insulation from global supply chain disruptions, coupled with the expansion of digitally deliverable services and knowledge-intensive activities. This is in line with (WTO, 2025[6]).
Within manufacturing and services, individual subsectors also display varying patterns (Figure 4). In manufacturing, subsectors such as computer, electronic and optical products, and motor vehicles showed modest increases in domestic value-added share between 2022 and 2024, while shares in machinery and equipment remained relatively stable. These sectoral differences reflect the heterogeneous nature of manufacturing global value chains (Mourougane et al., 2023[7]). Variations in services subsectors were less pronounced. Transport and storage, information and communication, and professional, scientific and technical activities all exhibited relatively stable patterns between 2022 and 2024, with only marginal increases. Recent analysis notes that services trade expanded particularly in digitally deliverable services and in business services including professional and technical services (OECD, 2025[8]).
OECD and large emerging markets, % of gross exports
Note: The large emerging economies include Brazil, China, India, Indonesia and South Africa.
Source: OECD TiVA 2026 nowcasting exercise.
Change in domestic services value-added share, percentage points
Source: OECD TiVA 2026 nowcasting exercise.
The continued growth of services in manufacturing
Between 2022 and 2024, experimental estimates indicate further “servicification” of manufacturing exports in many economies (Figure 5). On average, domestic and foreign services content in manufacturing exports both increased by around 0.3 percentage points over this period. A sizeable group of countries cluster in the upper-right quadrant of Figure 5, with both domestic and foreign services shares rising between 2022 and 2024. This group includes Australia, Canada, France, Japan, Korea, the Netherlands, Norway, Spain, and Sweden, as well as several smaller European economies (Belgium, Iceland, Lithuania, Luxembourg, Portugal, and Slovenia). For example, Japan’s domestic services content is estimated to have risen from about 19.8% to 20.5%, with foreign services content increasing from around 9.2% to 9.9%. Focusing on the most recent period, changes between 2023 and 2024 were generally modest suggesting that most adjustments occurred earlier, with 2024 largely consolidated earlier increases.
Taken together, these developments mark a shift compared with the pre-pandemic period, when rising services content in manufacturing exports was driven more strongly by foreign services. Between 2015 and 2019, domestic services shares in manufacturing exports declined slightly on average across OECD countries and large emerging economies, while foreign services shares increased by around 0.7 percentage points, pointing to a gradual offshoring of service tasks. The latest estimates for 2023-24, by contrast, suggest a more balanced dynamic, with both domestic and foreign services contributing to the rising services intensity of manufacturing exports.
Percentage points
Source: OECD TiVA 2026 nowcasting exercise.
Participation in global value chains can be viewed through two complementary lenses. Backward linkages describe the extent to which a country’s exports rely on foreign inputs – for instance, when firms import components from abroad to assemble and then export a finished product. By contrast, forward linkages describe how much of a country’s own production is used as inputs in another country’s exports – for example, a country exporting chemicals, parts, or business services that are then incorporated in another country’s exported goods and services. An economy is defined as “forward-oriented” when the value it adds to goods and services that are ultimately absorbed in another country’s final demand is larger than the foreign value it imports and embeds in its own exports. Conversely, “backward-oriented” economies rely more on imported inputs embodied in their exports than they do on supplying value added to the rest of the world (Banga, 2013[9]).
Based on experimental estimates for 2024, forward-oriented economies exhibited, on average, a foreign value-added share in exports of 25% and a domestic value added absorbed in foreign final demand of around 33% (Figure 6). Backward-oriented economies showed the opposite pattern, with foreign value-added shares closer to 35% and domestic value-added absorbed in foreign final demand around 30%. Small, highly open economies such as Luxembourg, Ireland, the Slovak Republic, and Hungary typically combined high foreign input use with strong forward linkages, though in different configurations. For example, Luxembourg recorded exceptionally high foreign value added (around 65%) alongside strong forward participation (just above 60%), while Ireland combined very high forward absorption (around 75%) with elevated foreign value added (around 42%), reflecting its role as a hub for multinational production and services.
Foreign value-added shares in exports compared to domestic value added absorbed in final demand, percentage points
Note: Backward participation (foreign value added share in exports) is plotted against forward participation (the percentage of total domestic value added absorbed in foreign final demand), and weighting by the degree of participation into global value chains (as measured by the sum of both indicators).
Source: OECD TiVA 2026 nowcasting exercise.
Large economies also occupied varied positions within global value chains. In 2024, Germany and the United States are estimated to be on the forward-oriented side of the 45-degree line in Figure 6, with relatively moderate foreign content shares (around 22% and 8%, respectively) and non-negligible domestic value added destined for foreign final demand (around 31% and 9%). Meanwhile, China, India and Korea remained somewhat backward oriented, relying slightly more on foreign inputs than on supplying value added to foreign final demand, as their foreign value added shares slightly exceeded their forward participation experimental estimates.
Overall, despite small increases in domestic value-added shares, participation in global value chains remained structurally high in both directions. Many economies that source significant foreign inputs still depended heavily on foreign demand to absorb their value added, a pattern consistent with earlier OECD analysis showing that global value chains participation is multi-dimensional, allowing countries to reduce specific dependencies without retreating from international production networks (Kano, Tsang and Yeung, 2020[10]).
Domestic value-added share (or content) of gross exports by industry i to partner region p represents the exported value added that has been generated anywhere in the domestic economy. This is an intensity measure and reflects how much value added, generated anywhere in the domestic economy, is embodied in total gross exports by industry.
Foreign value-added share (or content) of gross exports captures the value of imported intermediate goods and serves that are embodied in a domestic industry’s exports. The value added can come from any foreign industry upstream in the production chain. This is an intensity measure, often referred to as import content of exports and considered as a measure of backward linkages in analyses of GVCs. It reflects how much value added, generated abroad, is embodied in total gross exports by industry.
Domestic services share (or content) of gross exports includes value added originating from upstream domestic services industries that is embodied in exports (including exports of final services). This indicator is often used to measure services content embodied in manufacturing exports, to capture the rising importance of services integration in manufacturing production and exports.
Foreign services share (or content) of gross exports measures the share of value added originating from foreign services industries that is embodied in a domestic industry’s gross exports. It captures imported services used directly and indirectly in the production of exported goods and services and is often interpreted as the foreign services content of exports.
Share of domestic value added embodied in foreign final demand measures, for each industry, the share of its domestic value added that is ultimately absorbed in foreign final demand (consumption and investment) rather than in domestic final demand. It is commonly used as an indicator of an industry’s or economy’s reliance on foreign final demand and of its forward (upstream) participation in global value chains.
Note: Definitions are taken from Guide to OECD Trade in Value Added (TiVA) Indicators, 2025 edition, OECD 2025.
The 2024 TiVA nowcasting exercise provides an opportunity to assess how well machine-learning-based projections can anticipate trade in value added patterns in the absence of timely official statistics (Insights on engagement in global value chains in 2023). Using historical data to estimate 2022-23 TiVA indicators, and validating the procedure against newly released 2022 results, delivers a realistic benchmark of predictive accuracy.
Overall accuracy for the headline measures is shown to be strong (Figure 7). For the economy-wide indicators averaged across all TiVA measures, the mean absolute error is relatively modest, with the world average (excluding outliers) close to zero. This level of precision compares favourably with international nowcasting standards and is well within the tolerance commonly acceptable for economic diagnostics. It also reflects the relatively stable macro-trade fundamentals that machine learning models can capture with reasonable reliability.
Manufacturing shows somewhat larger errors than the economy-wide average, reflecting the sector’s greater sensitivity to supply chain disruptions and intermediate input volatility. Services, by comparison, exhibit more balanced errors around the world average, though country-specific variation remains. These sectoral patterns suggest that manufacturing value chains, with their complex intermediate linkages and just-in-time production networks, pose greater nowcasting challenges than services activities.
Average error across all TiVA indicators for economy-wide, percentage points
Note: The dashed line shows world average; dotted line shows cross-country average excluding top 25% outliers (faded bars). Red bars indicate overestimation (positive error); blue bars indicate underestimation (negative error).
Source: OECD TiVA 2024 and 2026 nowcasts.
Open economies that are typically more volatile, tend to exhibit larger nowcast errors, reflecting their greater exposure to idiosyncratic shocks, partner-specific trade dynamics, international investment, and sectoral concentration that are harder to capture with aggregate indicators. Denmark stands out with particularly large errors across sectors, likely reflecting its specialisation in maritime services, pharmaceuticals, and energy-related activities all of which experienced significant structural shifts during the validation period that proved difficult to anticipate from standard macroeconomic predictors.
Overall, the 2024 nowcasting exercise provides a credible early-warning tool for global value chain developments, while offering clear guidance on where methodological improvements would yield the largest gains in accuracy. The exercise highlights challenges in modelling complex manufacturing supply chains and services content, where both volatility and limited data availability constrain predictive accuracy. The concentration of larger errors among smaller, more specialised economies also offers a diagnostic for future refinement.
The exercise covers 41 economies and 24 sectors. Instead of reconstructing full input-output tables each year, the approach models the annual change in a small set of TiVA indicators (such as the domestic value-added share of exports) in a panel that pools all country-sector observations. On the right-hand side, the models use a large number of explanatory time series capturing domestic activity, trade, prices, labour markets, financial conditions, and global factors.
Several statistical and machine-learning models are estimated in parallel for each indicator: a simple autoregressive benchmark (using only the indicator’s own history), a fixed-effects linear regression, penalised regressions (ridge and lasso, which handle many correlated predictors and do variable selection), and a gradient-boosted trees (GBM) model to capture non-linear relationships and interactions. Key macro series capturing global macroeconomic developments like exports and value added are themselves projected forward so that the latest information can be fed into the TiVA models.
To select the optimal model for each country-sector-indicator combination, the paper uses time-respecting cross-validation: for each year in the past, it pretends to stand in that year, uses only data that would have been available then (taking publication lags into account), and forecasts one year ahead. The model with the lowest out-of-sample forecast error (root mean squared error) is selected and then used with the most recent macro and trade data to generate the nowcasts shown in the blog. Foreign value-added shares are finally obtained as a residual (one minus the nowcasted domestic share).
More information can be found in Mourougane, A. et al. (2023), “Nowcasting trade in value added indicators”, OECD Statistics Working Papers, No. 2023/03, OECD Publishing, Paris, https://doi.org/10.1787/00f8aff7-en.
The full database on TiVA indicators can be found here: Trade in value added | OECD.
References
[1] Andrenelli, A. et al. (2023), International Trade in the Wake of Multiple Shocks: OECD Global Trade Monitor.
[9] Banga, R. (2013), MEASURING VALUE IN GLOBAL VALUE CHAINS, UNCTAD.
[2] Crowe, D. and Ł. Rawdanowicz (2023), “Risks and opportunities of reshaping global value chains”, OECD Economics Department Working Papers, No. 1762, OECD Publishing, Paris, https://doi.org/10.1787/f758afe8-en.
[4] Jaax, A., S. Miroudot and E. Van Lieshout (2023), DEGLOBALISATION? THE REORGANISATION OF GLOBAL VALUE CHAINS IN A CHANGING WORLD Deglobalisation? The Reorganisation of GVCs in a Changing World.
[10] Kano, L., E. Tsang and H. Yeung (2020), Global value chains: A review of the multi-disciplinary literature, Palgrave Macmillan, https://doi.org/10.1057/s41267-020-00304-2.
[7] Mourougane, A. et al. (2023), “Nowcasting trade in value added indicators”, OECD Statistics Working Papers, No. 2023/03, OECD Publishing, Paris, https://doi.org/10.1787/00f8aff7-en.
[5] OECD (2025), OECD Supply Chain Resilience Review: Navigating Risks, OECD Publishing, Paris, https://doi.org/10.1787/94e3a8ea-en.
[8] OECD (2025), The OECD-WTO Balanced Trade in Services Database (BaTIS), OECD Publishing, Paris, https://doi.org/10.1787/c321a7a7-en.
[3] Schwellnus, C., A. Haramboure and L. Samek (2023), “Policies to strengthen the resilience of global value chains: Empirical evidence from the COVID-19 shock”, OECD Science, Technology and Industry Policy Papers, No. 141, OECD Publishing, Paris, https://doi.org/10.1787/fd82abd4-en.
[6] WTO, U. (2025), Recent Developments in Global Value Chains: Rewiring GVCs in a Changing Global Economy, WTO, UIBE, ADB, ID-JETRO, WEF.
Contact
Tom Arend (Statistics and Data Directorate, Trade and Productivity Statistics Division), tom.arend@oecd.org.