Climate Justice and Sectoral Responsibility for Energy-Related Carbon Emissions in Thailand: An Inter-Sectoral EEIO Framework

Agricultural production remains highly climate-sensitive, relying on rainfall patterns, temperature stability and ecosystem services that are increasingly disrupted by intensified droughts, floods and heat stress [1-3]. This escalating environmental volatility necessitates a systemic shift toward climate-resilient agriculture, a transition that is fundamentally predicated on the mobilization of significant capital investment. Substantial investment is required to bridge the widening adaptation gap, specifically for the development and deployment of climate infrastructure, such as advanced irrigation systems and flood-resistant storage facilities.

Grounded in climate justice, such financing mechanisms should extend beyond conventional public budgeting by incorporating redistributive instruments that channel resources from carbon high-emitting sectors to vulnerable agricultural systems [4, 5]. Under a carbon accountability approach, funds can be systematically collected from sectors ranked by their carbon intensity, such as energy, transportation, and heavy industry, and redirected toward smallholder farmers and climate-exposed rural communities [6, 7]. This approach internalizes environmental externalities while enhancing equity, ensuring that those contributing most to emissions bear a proportionate share of adaptation costs. In doing so, climate finance becomes both a corrective and enabling mechanism, strengthening agricultural resilience while advancing distributive justice across sectors.

Because production systems are deeply interconnected, sector specific carbon policies may generate economy wide adjustments [3, 8, 9]. To address this challenge, we therefore employ an environmentally extended input–output (EEIO) analysis framework to examine the economy wide and distributive implications of carbon taxation and sectoral emission responsibility in Thailand. We integrate sectoral greenhouse gas accounts into national input–output tables and attempt to trace both direct and embodied emissions across production and consumption networks.

In Thailand, recurrent El Niño–Southern Oscillation variability amplifies hydrological extremes, directly constraining crop yields, livestock productivity and water availability [10, 11]. Especially, related impacts on Thailand’s agricultural sector are projected to result in economic losses of THB 17,912–83,826 million per year between 2011 and 2045 owing to variability in temperature and rainfall, extreme events, and associated yield declines [12].

This context positions Thailand as a compelling case for examining climate justice–oriented financing in the agricultural sector. The country combines high climate exposure with structural dependence on smallholder farming, alongside uneven carbon emission contributions across sectors such as energy, transport, and industry. These conditions create a clear empirical basis for analyzing the inter-sectoral linkages of carbon emissions throughout the Thai economy. By tracing how emissions flow between high-emitting sectors and climate-vulnerable agriculture, this study establishes a structural framework to identify key leverage points for policy intervention. Ultimately, this enables policymakers to design targeted redistribution mechanisms and supportive policies that allocate resources from heavy emitters toward bolstering the climate-threatened agricultural sector.

3.1 Environmentally extended input–output

EEIO analysis links economic flows between sectors with environmental flows (emissions, resource use) to show how final demand drives environmental pressures along supply chains [13]. In this study, we applied this framework to quantify sectoral carbon emissions and examine the distribution of environmental responsibility across production systems, as demonstrated in prior studies on carbon management [14-16]. By coupling national input–output tables with environmental accounts, the model captures both direct emissions and those embodied in supply chains and establishes a rigorous empirical basis for assessing redistributive policy instruments, including carbon taxation and revenue recycling, within a governance framework that seeks to balance allocative efficiency with distributive equity.

Relative to the Agriculture, Forestry and Other Land Use (AFOLU) literature, which typically emphasizes land-use change, methane abatement, and on-farm efficiency [17], this approach reorients attention toward upstream energy inputs embedded in agricultural production. Rather than treating agriculture solely as a direct biological emitter, we employ an EEIO framework to reveal how a substantial share of agricultural embodied emissions originates in electricity generation, fertilizer production, petroleum refining, and transport services. This extends AFOLU mitigation pathways by identifying systemic leverage points beyond farm-level practices. Furthermore, compared with existing EEIO studies that focus primarily on consumption-based accounting or trade-embedded emissions [18], our integration of climate justice principles extends the analytical scope by examining structural transmission mechanisms. We link forward linkage centrality with emission intensity, arguing that responsibility is tied not only to the volume of emissions but also to network influence within the production system. By combining linkage analysis with scenario-based emission coefficient adjustments, this study moves EEIO applications from a static attribution approach toward a policy-relevant structural transformation analysis in emerging economies such as Thailand.

It is important, however, to clarify the accounting boundary. Our present EEIO specification primarily captures energy-related CO₂ emissions associated with fossil fuel combustion across sectors. Biological non-CO₂ emissions, most notably methane (CH₄) from rice cultivation and enteric fermentation in livestock, are not fully represented in the energy-based satellite accounts. Inclusion of these biological emissions would likely increase the absolute emission levels attributed to certain agricultural sectors, particularly paddy and livestock. Nonetheless, because the structural centrality of upstream electricity, fossil energy supply, and carbon-intensive materials arises from their strong forward linkages and economy-wide input provision, incorporating CH₄ would alter the magnitude of sectoral totals but not necessarily overturn the observed structural concentration of embodied carbon in energy-anchored industries. Accordingly, the results should be interpreted as reflecting energy-related carbon responsibility within the production network, while acknowledging that a comprehensive greenhouse gas assessment would require the integration of biological emissions inventories into the satellite framework.

Let $A$ denote the matrix of technical coefficients derived from the national input–output table, where each element represents the intermediate input required from sector $i$ to produce one unit of output in sector$~j$. Total sectoral output $X$ is determined by final demand $Y$ according to the standard Leontief formulation, as illustrated in Eq. (1):

$X={{\left( I-A \right)}^{-1}}Y$        (1)

where, ${{\left( I-A \right)}^{-1}}$ is the Leontief inverse matrix capturing both direct and indirect production requirements. Let $L={{\left( I-A \right)}^{-1}}$, then backward linkage or output multiplier of sector $j~\left( B{{W}_{j}} \right)$ is expressed in Eq. (2):

$B{{W}_{j}}=\underset{i}{\mathop \sum }\,{{L}_{ij}}$          (2)

and forward linkage can be noted in Eq. (3):

$F{{W}_{i}}=\underset{j}{\mathop \sum }\,{{L}_{ij}}$          (3)

Backward linkage captures the induced demand for upstream inputs generated by sectoral expansion, with the growing sector operating as a purchaser that stimulates activity across its supply chain. In contrast, forward linkage describes the transmission of output expansion from a given sector to downstream industries that use its products as intermediate inputs, whereby the originating sector functions as a supplier within the production network

To incorporate environmental dimensions, a vector of direct emission coefficients or intensities $\left( f \right)$ is constructed, representing carbon dioxide emissions per unit of sectoral output. Total embodied emissions associated with final demand $\left( E \right)$ are then estimated in Eq. (4) as:

$E=f{{\left( I-A \right)}^{-1}}Y$          (4)

where,

${{f}_{i}}=\frac{{{F}_{i,k}}}{{{x}_{i}}}$ =$~{{F}_{k}}{{\hat{x}}^{-1}}$

${{F}_{i,k}}~$= Total direct CO₂ emissions from sector$~i$ using energy type $k$

${{F}_{k}}$ = Column vector of sectoral emissions using energy type $k$

${{x}_{i}}$ = Total gross output of sector $i$

${{\hat{x}}^{-1}}~$= Diagonal matrix of inverse sectoral outputs

This formulation decomposes emissions into production- and consumption-based accounts, revealing how carbon-intensive sectors diffuse environmental burdens across supply chains. The relationship can be expressed as: emission intensities $\left( f \right)$ provide the basis for estimating direct emissions (${{F}_{i,k}}$), while the input–output structure extends these direct emissions across inter-sectoral supply chains to estimate embodied emissions $\left( E \right)$. Thus, direct emissions represent the emissions occurring within a sector, whereas embodied emissions capture the broader carbon footprint attributable to final demand, including emissions generated indirectly through the sector's purchases of intermediate inputs.

Therefore, sectors with strong forward linkages, such as electricity and petroleum refining, propagate emissions systemically because their outputs serve as essential inputs for multiple downstream activities [19, 20]. Structural position therefore matters alongside emission magnitude: a sector’s network centrality determines how widely its carbon is embodied throughout the economy [19, 21]. Responsibility, in this framework, is thus tied not only to the volume of emissions generated, but also to the sector’s capacity to transmit carbon through the production system.

3.2 Scenario analysis

Within an EEIO framework, a justice-oriented shock is introduced as an exogenous change. Therefore, in this study, we propose a simulation to explore the possibility of reallocating funds collected from sectors with high carbon emissions to support agriculture, which produces lower emissions, based on the Polluter Pays Principle.

Beginning with general assumption, climate damage reduces agricultural productivity. In input–output (I-O) terms, lower productivity means higher input requirements per unit output. This is modeled as an increase in agricultural column coefficients in matrix $A$, indicating in Eq. (5):

$A_{i,j}^{Shock}={{A}_{ij}}\left( 1+\theta  \right)$          (5)

where, $\theta $ represents climate-induced inefficiency (unit: %). Then the new output is shown as Eq. (6):

${{X}^{Shock}}={{\left( I-{{A}^{Shock}} \right)}^{-1}}Y$          (6)

Eq. (6) captures how climate stress amplifies backward linkages and increases embodied input dependence. In EEIO terms, resilience reduces technical coefficients (efficiency improvement) and/or shifts input structure away from vulnerable inputs. Therefore, when we propose climate funds (collected from high-emission sectors) finance to increase the resilience of agricultural sectors, the coefficient matrix of the agriculture can be defined as Eq. (7):

$A_{i,j}^{Res}={{A}_{ij}}\left( 1-\tau  \right)$          (7)

where, $\tau $ represents efficiency gain from resilience investment (unit: %). Then the new output is shown as Eq. (8):

${{X}^{Res}}={{\left( I-{{A}^{Res}} \right)}^{-1}}Y$          (8)

Then we decompose the effect as:

Output preservation ratio$~\left( OPR \right)=1-\frac{{{X}^{Res,agri}}}{{{X}^{Shock,agri}}}$          (9)

and change in carbon embodied emissions:

$E=f[{{\left( I-{{A}^{Res}} \right)}^{-1}}-{{\left( I-{{A}^{Shock}} \right)}^{-1}}]Y$          (10)

Eq. (9) captures the structural efficiency improvements in agriculture, illustrating how resilience financing reduces the volume of embodied emissions relative to the shock condition, while also decreasing systemic input amplification within the production network. Eq. (10), in turn, reflects the change in embodied emissions resulting from the reallocation of funds under different resilience financing schemes.

We propose that the reallocated funds must be invested in climate-resilience activities in agricultural sectors where the marginal efficiency gains are projected to exceed the marginal losses in agricultural productivity. Therefore, the value of $\tau $ should be higher than $\theta $. Table 1 demonstrates the baseline assessment.

Table 1. Baseline assessment of reallocating funds from key polluters to enhance agricultural climate resilience

Source: Developed by authors.

3.3 Data

This study simplified Thailand’s official input–output table from the Office of the National Economic and Social Development Council (NESDC) into 30 sectors. We grouped these sectors based on the production traits and economic activities of the original 180-sector table. This process keeps detailed information on key agricultural and high-emitting industrial sectors while keeping the model easy to calculate. The final 30 sectors include 12 agricultural and 18 non-agricultural activities. Details of the mapping process can be downloaded at the supplementary material.

We employed the 2015 I–O table rather than the 2021 edition to avoid structural distortions associated with the COVID-19 pandemic. The pandemic generated abrupt disruptions in production, consumption, and trade, producing atypical intersectoral linkages and temporary demand contractions that do not reflect the economy’s underlying structural relationships. Incorporating crisis-period data would risk conflating transitory shocks with long-run production structures, thereby biasing multiplier and linkage estimates.

In terms of carbon emission, we obtained data from Global Trade Analysis Project (GTAP) 12-Satellite Data and Utilities Year (2017 reference year) [22]. Emission coefficients from GTAP were used to calculate the sectoral carbon emission by multiplying with the output level reported in Thailand’s 2015 NESDC input–output table. To reconcile the two datasets, total emissions from all production sectors were scaled to match Thailand’s official 2015 greenhouse gas (GHG) emission level reported in World Resources Institute [23]. In this balancing adjustment, we also assume that production technologies, energy structures, and carbon intensities remain relatively stable over short timeframes, ensuring structural consistency between the emission satellite data and the national economic framework. All financial values will be reported in US dollars (USD), with local currency amounts converted at an exchange rate of THB 34.25 per USD, based on the 2015 annual average exchange rate [24].

5.1 Empirical discussion

The empirical results reveal a highly asymmetric carbon structure in which embodied emissions are concentrated in a small group of upstream energy and industrial sectors. In absolute terms, Electricity records the highest total emissions. Electricity records the highest total emissions because power generation is an intermediate input to nearly all sectors [25, 26]. Consistent with Muangthai et al. [25], electricity generation shows one of the highest forward linkage indices, explicitly indicating that it is “a vital input to other industries” and a key basis of industrial activity. In addition, transportation is strongly petroleum-based, confirming that embodied emissions in Thailand driven by final demand are structurally anchored in fossil-fuel systems. This figure reflects the direct combustion of liquid fuels in freight and passenger mobility, while pipeline activities rely on sustained oil and natural gas throughput. Because logistics and distribution networks underpin both production and final consumption, emissions from these sectors increase in close alignment with expanding trade, services, and household demand [27-29]. This reinforces their role as key transmission channels of embodied carbon across the economy.

By contrast, agricultural sectors display comparatively modest emission levels, consistent with existing studies showing that most crop-based agricultural activities are less energy-intensive per unit of output than heavy industry [30-32]. Energy and industry are dominated by CO₂ emissions from fossil fuel combustion, a long-lived stock pollutant that accumulates in the atmosphere [33, 34], but agriculture is dominated by methane (CH₄) and nitrous oxide (N₂O) emissions from enteric fermentation, manure, agricultural soils, and rice cultivation, with a relatively smaller share of fossil-fuel-based CO₂ [35, 36]. When all gases are converted into CO₂‑equivalents, the high fossil‑CO₂ volumes from upstream energy and industry overshadow non‑CO₂ gases from agriculture, making the agricultural sector’s share look smaller even where it is a leading CH₄/N₂O source [33, 36, 37].

A redistributive design for climate investment is essential. Our scenario analysis supports this approach, emphasizing a concept of climate justice grounded in differentiated responsibility, structural capability, and vulnerability. Resilience investment generates the largest efficiency gains in climate-vulnerable agricultural sectors, particularly rain-fed crops. Existing studies show that smallholder, rain-fed crops in Africa, Asia, and Latin America are highly sensitive to drought, floods, and heat, leading to substantial yield and income losses. Because baseline productivity and resilience are low, investments with a redistributive design that target sectors involved in their technical production, such as water management and infrastructure, can generate disproportionately large improvements in performance and resilience. In addition, embodied emission reductions are amplified in structurally central and carbon-intensive sectors, with most notably electricity and chemical production. Although investment is directed toward agriculture, input–output linkages transmit efficiency gains upstream, reflecting their strong backward linkages. This confirms that responsibility should not be assigned solely on the basis of direct emissions, but also according to their structural influence within the production network [20, 21].

In summary, the results demonstrate a quantitative mismatch between carbon responsibility and structural influence. High-forward-linkage sectors combine high emission intensity with strong transmission capacity, making them central drivers of embodied emissions. Agriculture, by comparison, contributes relatively low direct emissions and exhibits weaker forward influence, yet remains highly climate-sensitive. The transfer of responsibility from high-emission sectors to the agricultural sector is therefore required.

5.2 Policy discussion

By quantifying total embodied emissions driven by final demand, this study constructs illustrative structural scenarios to map how downstream consumption patterns generate upstream carbon loads across the supply chain (Figure 4). Rather than serving as empirical forecasts of fiscal mechanisms or specific tax instruments, these scenarios provide a baseline target map for policymakers. By identifying which non-agricultural sectors account for the highest carbon intensity linked to agricultural demand, this approach establishes a structural rationale for policy design—offering an empirical foundation to evaluate how mitigation efforts and supportive resource allocation could be strategically targeted across interconnected sectors.

Figure 4 synthesizes these structural findings into a conceptual policy framework for climate justice in agriculture, drawing on the rationale of the Polluter Pays Principle. By linking sectoral emission intensities identified in the EEIO model to overarching policy design, this framework illustrates how carbon-pricing mechanisms could conceptually assign greater responsibility to emission-intensive sectors, such as fossil energy, heavy manufacturing, and petrochemicals. Rather than simulating specific tax rates or fiscal revenue streams, this conceptual model highlights how revenues from high-emitting supply chains could strategically feed into a dedicated climate fund to support climate-vulnerable, low-emitting sectors like agriculture.

tu_pian_4.png

Targeted allocation to on-farm adaptation such as, clean technology deployment, resilient infrastructure and risk transfer instruments strengthens the adaptive capacity of farming systems that experience disproportionate climate exposure despite modest per unit emissions [17]. When revenue recycling is structured progressively, carbon pricing can mitigate regressive effects and enhance policy legitimacy [38, 39]. Embedded within a consumption-based accounting framework [40], this redistributive mechanism aligns mitigation obligations with emission drivers while directing financial flows toward sectors characterized by high livelihood sensitivity, thereby advancing climate justice through accountable public finance.

A climate justice scenario reallocates responsibility by tracing embodied emissions across the production network. Consumption-based accounting shows that high-revenue, carbon-intensive sectors drive disproportionate upstream emissions [41]. Linking these sectors to climate-vulnerable agriculture, such as rain-fed Paddy, Maize, and Sugarcane, aligns responsibility with structural drivers, while directing fiscal support toward sectors with limited adaptive capacity. Mitigation and resilience thus become mutually reinforcing. Structural transformation complements this approach. Shifting final demand from coal-intensive materials toward higher value-added manufacturing and services lowers embodied emissions through compositional change [42, 43]. Embedded in just transition policies, such shifts redistribute opportunity and prevent carbon lock-in [44].

Furthermore, consumption reallocation away from electricity-intensive materials, petrochemicals, construction, and transport targets sectors with strong forward linkages, where small demand changes generate amplified upstream emission reductions [45, 46]. Evidence from China indicates that dominant emission pathways are concentrated along a limited number of structurally embedded supply chain routes [46, 47]. Structural path and decomposition analyses show that trajectories such as non-metallic minerals to construction to capital formation, and electricity to intermediate production sectors to final consumption, account for a substantial share of embodied carbon dioxide and methane emissions. These findings highlight the centrality of investment driven construction demand and electricity intensive intermediate inputs in amplifying upstream emission flows. Accordingly, moderating capital formation and construction related expenditure can suppress emissions in core supplying sectors, particularly energy generation and basic materials, generating system-wide mitigation effects that exceed the immediate reduction in final demand.

The energy transition scenario should place explicit emphasis on upstream decarbonization by reducing emission intensities in electricity generation, petroleum refining and natural gas supply through renewable substitution and efficiency gains. These interventions are especially impactful in sectors with strong forward linkages including with energy supply chains that influence downstream sectors [48]. In oil and gas systems, the transitional options include process efficiency improvements, carbon capture and storage (CCS), fuel switching and the integration of renewable energy into refinery and upstream operations. These measures can deliver near-to medium-term emission reductions in contexts where fossil fuels remain embedded in the energy mix [49-51]. By lowering emission coefficients in core supplying sectors, the energy transition scenario generates cascading reductions in embodied emissions throughout supply chains, thereby reinforcing both mitigation effectiveness and distributive equity. These strategies align with evidence from emerging economies where electricity and petroleum sectors function as structural emission hubs. Studies of China and India show that decarbonizing power generation yields economy-wide mitigation effects disproportionate to the sector’s direct emission share because of its pervasive forward linkages [52, 53]. Similar dynamics are observed in Southeast Asia (see Discussion).

This study was supported by Thammasat University Research Fund, Contract No. TUFT 003/2568.