Marginal abatement cost curves for carbon-neutral transition in Thailand’s agri-food processing industry: quantitative pathways toward bio-circular-green economy alignment

Abstract

Background:

Thailand’s agri-food processing sector is a significant source of industrial greenhouse gas (GHG) emissions, yet the cost-effectiveness of sector-level decarbonisation options remains unquantified in the peer-reviewed literature. This evidence gap hinders evidence-based policy design under Thailand’s Bio-Circular-Green (BCG) Economy strategy and its Nationally Determined Contribution (NDC) target of 33.3% GHG reduction by 2030.

Methods:

This study constructs sector-specific Marginal Abatement Cost (MAC) curves for five major Thai agri-food processing sub-sectors (sugarcane processing, cassava starch manufacturing, rice milling, canned fruit, and frozen seafood processing) using exclusively secondary quantitative data from national energy audit reports, GHG inventories, and IPCC emission factors. Eight abatement measures are evaluated per sub-sector. Parameter uncertainty is bounded through Monte Carlo simulation. A composite BCG Alignment Index (BCG-AI) is introduced to score each sub-sector across the Bio, Circular, and Green dimensions of the BCG framework.

Results:

Total Scope 1 and 2 GHG emissions from the five sub-sectors are estimated at 8,737 ktCO₂e per year. MAC values range from −14.2 USD per tonne CO₂ equivalent for biomass cogeneration in sugarcane processing to +42.5 USD per tonne CO₂ equivalent for cold-chain electrification in frozen seafood. The Aligned Transition scenario yields an estimated abatement of 3,399 ktCO₂e (90% uncertainty interval: 2,780–3,910 ktCO₂e), exceeding an analytical proxy sector-level NDC target of approximately 2,910 ktCO₂e. No sub-sector achieves BCG Advanced status (BCG-AI > 70).

Conclusion:

This study presents one of the first sector-wide MAC curve analyses for Thailand’s agri-food processing industry published in the international peer-reviewed literature. NDC-consistent abatement is technically achievable at negative or near-zero net cost through biogas recovery and biomass cogeneration, and the findings deliver an actionable policy matrix for BCG industrial strategy and a baseline emissions-intensity dataset that Thai food exporters can use for voluntary carbon footprint disclosure as international carbon trade governance evolves.

1 Introduction

The global food system accounts for approximately 21–37% of total anthropogenic GHG emissions (IPCC, 2022; Crippa et al., 2021). Food systems are therefore a critical target for decarbonization policy, and the transition towards net zero in food and land-use systems presents both technical and governance challenges that require integrated policy frameworks across production, processing, and distribution stages (Reay et al., 2020). Michel et al. (2024) demonstrated that food processing delivers substantial nutritional and economic benefits throughout global value chains yet simultaneously generates energy-intensive emissions that are inconsistently tracked across national inventories.

Pearson et al. (2023) further emphasized that effective decarbonization of food systems requires transparent, standardized GHG measurement across supply chain nodes, a prerequisite that this study addresses through sector-specific MAC curve construction for Thai agri-food processing.

Thailand occupies a uniquely strategic position in the global food system as one of the world’s foremost agri-food exporters, ranking among the top global producers and exporters of cassava starch, cane sugar, rice, canned pineapple, and frozen seafood, with food processing export revenues exceeding 1 trillion THB in 2019 (BCG Thailand, 2021). Thailand’s national GHG inventory identifies the energy sector as the largest contributor to national emissions at 65.89% (254,307 ktCO₂eq), with industrial processes and products contributing an additional 10.50% (40,527 ktCO₂eq) (UNDP Thailand, 2023).

Thailand’s Second Nationally Determined Contribution targets an unconditional GHG reduction of 33.3% (184.8 MtCO₂eq) below the business-as-usual scenario by 2030, with carbon neutrality by 2050 and net-zero emissions by 2065 (Climate Action Tracker, 2025; CCPI, 2025). These targets are embedded within the BCG Economy Model, adopted as Thailand’s national development strategy from 2021 to 2026 (BCG Thailand, 2021; Jaroenkietkajorn et al., 2024).

MAC curves have been applied to agricultural sectors in Ireland (Teagasc, 2023) and the United States (USDA, 2023), but the methodology has not been widely extended to Southeast Asian agri-food processing. This study extends the MAC framework to five Thai sub-sectors and integrates it with Thailand’s BCG Economy Model a combination that, to the authors’ knowledge, has not previously been operationalized in the peer-reviewed literature.

This study addresses the gap through four objectives: (1) to establish a quantitative GHG emissions baseline for five Thai agri-food processing sub-sectors; (2) to construct sector-specific MAC curves ranking abatement measures by cost-effectiveness; (3) to develop a BCG Alignment Index scoring sub-sector readiness across the Bio, Circular, and Green dimensions; and (4) to assess the feasibility of achieving Thailand’s NDC 2030 targets under three decarbonization scenarios.

2 Literature review

2.1 Global food system emissions and the processing sector

Crippa et al. (2021) showed that food systems account for approximately 34% of global anthropogenic GHG emissions. Jones et al. (2023) demonstrated that historical emissions from food and land-use systems have been systematically underestimated in national inventories, with Southeast Asian nations showing particularly large upward revisions when agricultural waste gases are included. Feng et al. (2023) showed through continental-scale comparison that Asian food production systems exhibit higher energy-related emission intensities in the processing stage than European counterparts, reflecting greater reliance on fossil fuel-based industrial energy. Meng et al. (2024) identified that economic growth, energy structure, and production scale are the dominant drivers of agricultural carbon emissions in rapidly developing economies, findings directly applicable to the Thai food processing context examined here.

The trade dimension of food system emissions is particularly relevant for export-oriented Thai processors. Sandström et al. (2018) demonstrated that international trade redistributes the GHG footprints of food consumption across national boundaries, with food imports from developing countries often carrying higher embedded carbon intensities. Burke et al. (2023) quantified farm-to-fork GHG emissions across five dietary patterns in Europe and North America, establishing the empirical basis for trade-linked carbon adjustment mechanisms such as the EU Carbon Border Adjustment Mechanism (CBAM). Michel et al. (2024) further showed that food processing’s energy-intensive emissions are inconsistently tracked across national inventories, limiting the evidence base for targeted mitigation policy.

2.2 MAC curve theory and applications

The Marginal Abatement Cost curve depicts the cost of reducing one additional unit of GHG emissions against the cumulative abatement potential achievable through successive mitigation measures (Senatla et al., 2013). The MAC curve methodology was popularized internationally by McKinsey and Company (2009), whose global GHG abatement cost curve became the foundational framework for sector-specific MAC analyses. The standard MAC definition formula is:

Two main approaches exist: the expert-based (bottom-up) approach, which models individual technologies with defined cost and potential parameters; and the model-based (top-down) approach using computable general equilibrium or energy systems models. Varga and Roeger (2022) employed multi-sector dynamic general equilibrium modelling to assess EU climate targets, while Fragkos and Fragkiadakis (2022) highlighted that bottom-up approaches retain superior granularity for sector-specific technology analysis. This study employs the expert-based approach, following Teagasc (2023) and USDA (2023). The operational MAC formula presented later is an expanded form of this definition, where the cost difference is decomposed into annualized capital and operating cost components; the two formulations are mathematically equivalent.

2.3 Thailand’s bio-circular-green (BCG) economy framework

Thailand’s BCG Economy Model was formally adopted as the country’s national development strategy in 2021, building on bioeconomy, circular economy, and green economy principles (BCG Thailand, 2021; APEC, 2022). Jaroenkietkajorn et al. (2024) identified that financing constraints, technology transfer gaps, and insufficient monitoring frameworks are the primary BCG implementation barriers. Smol et al. (2020) and Tvaronavičienė (2024) both emphasized that genuine circular economy achievement requires simultaneous decarbonization of energy inputs. Fatimah et al. (2024) developed circular economy maturity assessment frameworks that share methodological features with the BCG Alignment Index introduced here. Mudzielwana (2025) highlighted that financial constraints, inadequate technical support, and fragmented policies are universal barriers to climate-smart food system implementation, consistent with Jaroenkietkajorn et al.’s (2024) assessment of the Thai BCG context.

2.4 Carbon emissions from Thai agri-food processing sub-sectors

2.4.1 Sugarcane processing

Yuttitham et al. (2011) reported a carbon footprint of 0.55 kgCO₂e/kg sugar from sugarcane cultivation and milling in eastern Thailand, with the milling stage contributing approximately 0.06 kgCO₂e/kg. Li et al. (2024) found that mills deploying bagasse-fired combined heat and power systems achieve carbon intensities 40–60% lower than those relying on external energy, directly supporting the high abatement potential estimated for biomass cogeneration in this study.

2.4.2 Cassava starch manufacturing

Usubharatana and Phungrassami (2015) reported cassava starch carbon footprints of 130–572 kgCO₂e/ton across facilities in northeast Thailand. Hansupalak et al. (2016) demonstrated that biogas recovery from cassava wastewater reduces the carbon footprint by approximately 25–35%. Trakulvichean et al. (2025) together confirm a facility-level range of 130.9–572.3 kgCO₂e/ton with a median of approximately 468 kgCO₂e/ton across 12 plants, with electricity (33%) and fuel oil (48%) as the dominant emission sources.

2.4.3 Rice milling and other sub-sectors

Yodkhum et al. (2017) found that rice milling-stage emissions in Thailand are dominated by electricity consumption, with widespread husk utilisation contributing to relatively low carbon intensities. For canned fruit and frozen seafood, DEDE energy audit data provides sector-level benchmarks. Feng et al. (2023) identified that food processing sub-sectors with high refrigeration loads exhibit among the highest energy intensity globally. Meng et al. (2024) further showed that energy structure and production scale are the primary drivers of carbon intensity in agricultural processing industries.

2.5 Carbon pricing, trade policy, and the EU CBAM

Because the EU CBAM’s initial product scope does not yet include food products, references to CBAM throughout this manuscript are framed as forward-looking contextual motivation rather than a direct policy output of the present analysis; scope expansion is anticipated (Pollitt et al., 2020; Ruiz et al., 2023). Clora and Yu (2022) modelled 31 European decarbonization pathways, finding food and agricultural processing to be a logical future CBAM category. Schubert et al. (2026) demonstrated that border carbon adjustments are welfare-improving in second-best policy environments where domestic carbon pricing is incomplete. Vandyck et al. (2016) showed that NDC-consistent pathways in developing economies frequently require both domestic policy action and supportive international financing. Pearson et al. (2023) emphasized that transparent, standardized GHG data exchange across food supply chains is an essential prerequisite for the carbon accounting that any future border carbon adjustment regime would require from Thai food exporters.

Three specific research gaps are identified. First, no study has systematically compared baseline carbon intensities across multiple Thai food processing sub-sectors within a unified GHG accounting framework. Second, no MAC curve analysis has been published for Thailand’s agri-food processing sector in an internationally peer-reviewed journal. Third, no composite index for assessing BCG policy alignment at the sub-sector level exists, limiting monitoring of green industrial transformation. This study directly addresses all three gaps.

3 Materials and methods

3.1 System boundary and scope

This study applies a gate-to-gate system boundary for the processing stage of each sub-sector’s value chain, encompassing energy consumption, refrigerant use, and on-site wastewater treatment. GHG accounting follows the GHG Protocol Corporate Standard (Scope 1: direct emissions; Scope 2: purchased electricity). Five sub-sectors are analyzed: sugarcane processing, cassava starch manufacturing, rice milling, canned fruit processing, and frozen seafood processing, selected on the basis of economic significance, data availability, BCG relevance, and potential future CBAM exposure.

Scope 3 emissions, including upstream agricultural cultivation, fertiliser and pesticide manufacture, raw-material transport, packaging, and downstream distribution, are explicitly excluded from this gate-to-gate boundary. The decision to focus on Scope 1 + 2 reflects (i) the primary policy concern of the BCG industrial-processing pillar, which targets in-factory operations; (ii) the more reliable secondary-data coverage available for processing-stage emissions; and (iii) the requirement to keep the MAC analysis tractable across five sub-sectors. We acknowledge that for agri-food systems Scope 3 typically dominates total life-cycle emissions: indicative literature ranges suggest Scope 3 accounts for approximately 60–80% of life-cycle emissions in cassava starch (Hansupalak et al., 2016), 70–85% in canned tropical fruit, and 75–90% in frozen seafood (Feng et al., 2023), with raw-material cultivation typically the dominant contributor. Inclusion of Scope 3 would (i) substantially raise sub-sector total emissions; (ii) shift the relative ranking of sub-sectors towards seafood and tropical-fruit, whose upstream feed, water-use, and refrigerated-transport emissions are higher; and (iii) reduce the apparent share of total life-cycle abatement attributable to processing-stage MAC measures. The within-processing MAC ranking is, however, expected to remain robust because each measure is evaluated against its own gate-to-gate baseline. A full Scope 3 extension covering upstream cultivation and downstream cold-chain logistics is identified as priority follow-on work.

3.2 Data sources and coverage

This study relies exclusively on publicly accessible secondary data. No primary facility-level survey of Thai food processing GHG emissions exists in the peer-reviewed literature at sufficient sub-sector coverage to support cross-sector MAC analysis. The use of secondary data from authoritative government and international sources follows precedent established by Teagasc (2023) and USDA (2023). Table 1 provides a complete data inventory specifying source, document title, URL domain, and coverage for each data input.

Three coverage limitations follow: (a) DEDE audit reports provide sector-level aggregate energy intensity benchmarks rather than facility data, masking intra-sector heterogeneity; (b) TGO GHG inventory reports industrial emissions at two-digit ISIC level, requiring proportional disaggregation to derive sub-sector totals; and (c) cassava starch carbon intensity spans 130.9–572.3 kgCO₂e/ton across peer-reviewed studies (Trakulvichean et al., 2025; Usubharatana and Phungrassami, 2015), so the median value of 468 kgCO₂e/ton represents a sector-mean that may diverge from any individual facility. MAC values are therefore sector-representative central estimates rather than precise facility-level predictions. Sector-level aggregation may bias these values in two systematic ways: because mid-size and small facilities typically exhibit higher carbon intensities than the largest mills, using a sector-mean carbon intensity tends to under-state baseline emissions and over-state the cost-effectiveness of measures evaluated against that mean; and because abatement-measure CAPEX scales sub-linearly with facility size, aggregating CAPEX as a per-facility mean tends to under-state per-tonne abatement cost for small and medium enterprises. Both biases are partially captured by the parameter uncertainty intervals reported in the Results, but full resolution requires the stratified primary-data sample of at least 30 facilities per sub-sector that is recommended in the Limitations.

The per-facility CAPEX/OPEX values used in this study and the eligible-fleet scaling logic presented in Appendix implicitly assume broadly homogeneous facility size, energy mix, and operational profiles within each sub-sector a homogeneity assumption that warrants explicit acknowledgement. The assumption is least defensible for cassava starch, where the documented facility-level carbon-intensity range of 131–572 kgCO₂e/ton (Trakulvichean et al., 2025) spans more than a factor of four, and for rice milling, where commercial-mill capacity ranges from 10 to over 500 t/day. Under realistic intra-sector heterogeneity, aggregate sector-level MAC values are likely to (i) under-state per-tonne abatement cost for the smallest 30–40% of facilities, where CAPEX per unit output is higher and economies of scale in biogas digestion or cogeneration are not realised; (ii) over-state cost-effectiveness for measures whose technical performance scales with facility throughput; and (iii) under-represent regional clustering effects (e.g., lower transmission costs near sugar mills in Kanchanaburi versus dispersed cassava starch facilities in Nakhon Ratchasima). The Monte Carlo simulation captures part of this dispersion through the ±20% CAPEX variance, but full resolution requires a stratified primary-data sample of at least 30 facilities per sub-sector spanning small, medium, and large size tiers. The sector-level MAC results are accordingly presented as central estimates suitable for sector-strategic policy design rather than as facility-level investment-grade calculations, with the stratified primary-data sample identified as priority follow-on work.

3.3 GHG baseline calculation

Baseline GHG emissions (Scope 1 + 2) for each sub-sector are calculated using:

Where E_i is total GHG emissions for sub-sector i (tCO₂e/year); Activity Data_ij represents fuel consumption (TJ), electricity consumption (MWh), or refrigerant charge (kg); and Emission Factor_ij is the corresponding IPCC AR6 Tier 1 emission factor. The IEA Thailand grid emission factor of 0.4768 tCO₂/MWh (2022) is applied for Scope 2 calculations. The carbon intensity metric (kgCO₂e/ton of product) is derived by dividing total emissions by OIE annual production volume statistics. The arithmetic relationship CI × Production Volume = Total GHG (e.g., 55 kgCO₂e/ton × 95.0 Mt. = 5,225 ktCO₂e for sugarcane) is internally consistent across all five sub-sectors, as verified in Table 2.

3.4 MAC curve construction

Following Senatla et al. (2013), eight abatement measures are evaluated per sub-sector. The operational MAC formula is an expanded form of the standard definition above, where the cost difference is decomposed into annualised capital and operating cost components:

CAPEX and OPEX unit costs for each of the 15 abatement measures are provided in Table 3. These figures represent per-facility costs sourced from BOI investment databases, DEDE technology deployment reports, and IEA Southeast Asia renewable energy cost surveys. To derive sector-level inputs for the MAC formula, per-facility costs are multiplied by the estimated number of active facilities per sub-sector. The eligible-fleet sizes used here are 52 mills for sugarcane (active mills with crushing capacity above 5,000 t/day), 96 plants for cassava starch (installed capacity above 50 t/day), 720 commercial rice mills (above the 10 t/day threshold, applying a 60% large-mill eligibility filter), 25 plants for canned fruit (above 100 employees), and 70 plants for frozen seafood (above 50 t/day cold-storage capacity), derived from DEDE (2022) audit coverage, OIE Manufacturing Production Index records, and the BOI investment registry.

A measure-specific applicability factor (default 0.90 for retrofit measures and 1.00 for newly built solar PV) is then applied. Sector-level aggregate MAC is computed as the sum of fleet CAPEX times the capital recovery factor plus aggregate operating cost change, divided by aggregate abatement potential, with CRF = r(1 + r)^L / [(1 + r)^L − 1] at r = 8% and L = technology lifetime (Table 3). The full per-measure scaling worked example (Sugarcane biomass cogeneration), the eligible-fleet derivation logic, and the complete CAPEX–CRF–OPEX–abatement reconciliation table for all 15 measures are provided in Appendix and in the Supplementary material so that the calculations underlying Table 4 can be reproduced exactly. The resulting sector-level MAC values are reported in Table 4 and visualised in Figure 1.

Figure 1

To bound parameter uncertainty, a simulation-based Monte Carlo analysis (n = 10,000 iterations) is performed alongside the deterministic estimates. The simulation samples four input parameters: CAPEX ~ Normal(μ, 0.20 μ), OPEX ~ Normal(μ, 0.15 μ), emission factor ~ Normal(μ, 0.10 μ), and adoption rate ~ Triangular(min, mode, max), all truncated at zero where applicable.

The chosen variance ranges follow established precedent in techno-economic uncertainty analysis: the ±20% CAPEX coefficient of variation is consistent with the IEA (2023) renewable-energy cost-survey range for Southeast Asia and aligns with the conservative bound used in the original deterministic sensitivity analysis; the ±15% OPEX range reflects observed dispersion in BOI-DEDE energy-audit deployment records; and the ±10% emission-factor range corresponds to the Tier 1 uncertainty band reported in IPCC AR6 (2022).

Point estimates are accompanied by 90% uncertainty intervals derived from the empirical distribution of simulation outputs (5th and 95th percentiles), without recourse to Bayesian inference. The simulation also yields the rank-order stability of negative-MAC measures, defined as the proportion of runs in which each measure retains its cost-effectiveness ranking.

3.5 Abatement measures evaluated

Eight abatement measures are assessed across all five sub-sectors, adapted for sector-specific applicability:

Measure 1: Rooftop solar PV: Electricity substitution from grid to on-site solar. Capital cost: 650–850 USD/kWp (IEA, 2023).

Measure 2: Biogas recovery from wastewater process: Methane capture from anaerobic wastewater treatment. High applicability in cassava starch and sugarcane mills (Hansupalak et al., 2016).

Measure 3: Boiler efficiency upgrade: Replacement of legacy boilers with high-efficiency models, reducing fuel consumption 15–25%.

Measure 4: Variable speed drives (VSD) on motors: Retrofit of large motors (>30 kW), reducing electricity consumption 20–40%.

Measure 5: Biomass cogeneration from bagasse/husks: Combined heat and power from agricultural residues, displacing fossil fuel combustion (Li et al., 2024).

Measure 6: LED lighting and building energy management: Low-capital measure applicable across all sub-sectors. Typical payback period 2–4 years.

Measure 7: Refrigerant switching (R-22 to low-GWP alternatives): Replacement in cold-chain operations. High relevance to frozen seafood (Feng et al., 2023).

Measure 8: Renewable heat integration (solar thermal): Substitution of fossil fuel-fired process heat for washing, sterilization, and drying.

3.6 BCG alignment index

A composite BCG Alignment Index (BCG-AI) is constructed to score each sub-sector across the three BCG dimensions (BCG Thailand, 2021; Jaroenkietkajorn et al., 2024), informed by circular economy maturity frameworks (Fatimah et al., 2024) and critical success factor analyses (Sfakianaki, 2019). The BCG-AI is a first-generation, proof-of-concept index based on secondary proxy indicators; scores should be interpreted as indicative rankings rather than precise measurements. Mudzielwana (2025) similarly noted that composite sustainability indices derived from secondary data require empirical validation against primary facility-level assessments before guiding specific investment decisions.

The BCG-AI introduced in this study is positioned as a preliminary, proof-of-concept composite indicator. Its sub-dimension scores are derived from secondary proxy variables — residue-utilisation ratios, wastewater-recovery rates, renewable-energy share, and best-available-technology benchmark ratios and have not been empirically validated against (i) primary facility-level audits, (ii) expert-panel BCG-readiness assessments, or (iii) external benchmark composite indicators such as the OECD Green Growth Indicators or the Global Reporting Initiative G4 sector frameworks. To make the path to validation explicit, three concrete future-research pathways are proposed: (a) a Delphi study with 15–20 BCG-implementation experts drawn from MHESI, BOI, TGO (2023), academia, and industry associations, designed to elicit consensus weights for the three dimensions and to test the empirical defensibility of the 50/70 tier thresholds; (b) a primary facility-level survey of at least 30 plants per sub-sector, populating each sub-indicator with audit-grade data such that secondary-proxy bias can be quantified; and (c) Principal Component Analysis applied to a multi-country dataset (Thailand, Vietnam, Indonesia, the Philippines) to test whether the three-dimensional structure replicates outside Thailand. Until these validation steps are completed, BCG-AI scores should be interpreted as indicative rankings for policy-priority discussions rather than as inputs to specific investment-allocation, regulatory-tier, or fiscal-incentive decisions.

Equal weighting across the Bio, Circular, and Green dimensions is adopted as a transparent and theoretically neutral baseline, consistent with the BCG Thailand (2021) and APEC (2022) policy documents, neither of which specifies relative dimension weights, and following the precedent of first-generation composite sustainability indices (Smol et al., 2020; Fatimah et al., 2024) where empirical evidence is insufficient to defend asymmetric weights. The classification thresholds (above 70 = Advanced; 50–70 = Transitioning; below 50 = Emerging) are derived from the natural-breaks pattern observed across the five sub-sectors and align with the OECD (2024) tertile classification used in Thailand’s green-transition policy review.

To verify robustness, alternative weighting schemes Bio-heavy (50/25/25), Green-heavy (25/25/50), and Circular-heavy (25/50/25) were tested; tier classification is preserved for four of five sub-sectors, with only sugarcane shifting between Transitioning and the lower edge of Advanced under Bio-heavy weighting. Both the equal-weight default and the tertile thresholds are acknowledged as proof-of-concept choices that future work should refine through Analytic Hierarchy Process, Delphi expert elicitation, or Principal Component Analysis.

Bio dimension (B-score): Ratio of agricultural residue utilized to total generated; penetration of biogas/bioenergy in energy mix.

Circular dimension (C-score): Wastewater recovery rate, residue utilization rate, and material efficiency index (Smol et al., 2020).

Green dimension (G-score): Ratio of current carbon intensity to best-available-technology benchmark; share of renewable energy in energy mix (Tvaronavičienė, 2024).

BCG-AI = (B-score + C-score + G-score) / 3. Classification: BCG-AI > 70 = BCG Advanced; 50–70 = BCG Transitioning; < 50 = BCG Emerging. All five composite scores in Table 5 are verified by direct computation from the three-dimension scores.

3.7 Scenario design

Scenario 1: Business as Usual (BAU): No additional abatement measures are adopted; emissions grow at 1.5% per annum. This central rate is the 2018–2022 compound annual growth rate of DEDE-reported industrial energy use in food processing, integrating production-volume growth of approximately 2.0% per annum with autonomous efficiency gains of approximately 0.5% per annum.

Scenario 2: Moderate Transition: The four lowest-MAC measures are adopted by 2030 at a 60% adoption rate in large facilities (above 100 employees) and a 30% rate in small and medium enterprises. The 60/30 split is anchored to the historical 2015–2022 BOI-supported deployment record for renewable-energy and energy-efficiency tax incentives in Thai food processing, which shows 55–65% uptake among facilities with more than 100 staff and 25–35% uptake among smaller facilities, with the mid-points adopted as central estimates.

Scenario 3: Aligned Transition (BCG-NDC): All eight measures are adopted at technically feasible rates, requiring BOI investment incentives, carbon pricing instruments, and technical assistance (OECD, 2024; Fragkos and Fragkiadakis, 2022). The theoretical maximum abatement from the full set of 15 measure-sub-sector combinations at 100% adoption is 3,578 ktCO₂e. A central feasibility weighting of 95% is applied to reflect realistic adoption constraints, yielding the central estimate of 3,399 ktCO₂e (3,578 × 0.95 = 3,399.1, reported as 3,399).

Because this single weighting is the most consequential scenario assumption, sensitivity is reported across four feasibility values: 70% yields 2,505 ktCO₂e (the analytical proxy NDC target is not met), 80% yields 2,862 ktCO₂e (the proxy is marginally met), 90% yields 3,220 ktCO₂e (the proxy is exceeded), and 95% yields 3,399 ktCO₂e (the proxy is clearly exceeded). The qualitative conclusion that NDC-consistent abatement is technically achievable therefore holds for any feasibility value at or above 80%; below 80%, the proxy is missed without complementary policy levers.

The 95% feasibility weighting applied to the Aligned Transition scenario should be interpreted as an upper-bound technical-potential assumption rather than an empirical adoption forecast, and the framing is revised accordingly throughout this study. The 95% case represents a technical ceiling that would require an aggressive and well-financed implementation pathway combining mandatory carbon-intensity benchmarking, BOI Section 31/33 capital-cost subsidies, an operational T-VER aggregation mechanism for SMEs, and a domestic carbon price in the 15–25 USD/tCO₂e range. For policy-planning purposes, the 80–85% feasibility band is therefore recommended as the central operational assumption, broadly consistent with the historical 2015–2022 BOI deployment record for analogous renewable-energy and energy-efficiency incentives in Thai food processing (60–65% in large facilities, 30–35% in SMEs, weighted by the large-facility share of sectoral output). Under the 80% case (2,862 ktCO₂e), the analytical proxy NDC sector target of 2,910 ktCO₂e is marginally missed by approximately 48 ktCO₂e, and additional levers outside the eight-measure portfolio including process-heat heat-pump integration, demand-side manufacturing-process optimisation, and Scope 3 measures such as low-carbon logistics become necessary. Under the 70% case (2,505 ktCO₂e), the proxy is missed by approximately 14% and complementary supply-side and demand-side measures are indispensable. The headline conclusion that “NDC-consistent abatement is technically achievable” should therefore be read as conditional on a feasibility weighting at or above 80%, which itself depends on the policy and financing environment described above.

4 Results

4.1 Baseline carbon intensity by sub-sector

Table 2 presents the calculated Scope 1 + 2 baseline carbon intensity for each sub-sector. The cassava starch sub-sector exhibits the highest carbon intensity, with a median value of 468 kgCO₂e/ton (facility-level range: 131–572 kgCO₂e/ton; Trakulvichean et al., 2025), reflecting heterogeneity in facility scale, energy source mix, and wastewater management. Rice milling exhibits the lowest intensity (42 kgCO₂e/ton), reflecting widespread husk utilisation. Total GHG (8,737 ktCO₂e) is arithmetically verified: CI × Production Volume for each sub-sector sums exactly to the reported total.

To address concerns about the secondary-data basis of these sector-mean values, each sub-sector entry was cross-checked against the available plant-level peer-reviewed evidence: (i) the cassava starch median of 468 kgCO₂e/ton lies within the 131–572 kgCO₂e/ton facility-level range reported by Trakulvichean et al. (2025) across 12 plants in northeast and central Thailand, with the central value sitting in the upper-middle of the distribution where electricity-intensive starch-extraction units predominate; (ii) the sugarcane processing value of 55 kgCO₂e/ton is consistent with the milling-stage 0.06 kgCO₂e/kg sugar (= 60 kgCO₂e/ton) reported by Yuttitham et al. (2011) for three eastern Thailand mills, and falls within the 30–65 kgCO₂e/ton range reported by Li et al. (2024) for bagasse-CHP-equipped mills; (iii) the rice-milling value of 42 kgCO₂e/ton aligns with Yodkhum et al. (2017) for husk-fired Northern Thai mills. For canned tropical fruit and frozen seafood, no plant-level peer-reviewed Thai facility data is currently available, so the values for these two sub-sectors rest on DEDE sector-aggregate audit data alone; this validation gap is acknowledged and is a primary motivation for the stratified primary-data sample identified as priority follow-on work in Section 5.6. Within the available evidence base, the three validated sub-sectors fall within published facility-level ranges, supporting the use of these central estimates for sector-strategic MAC analysis.

4.2 MAC curve analysis by sub-sector

Table 4 presents MAC values and abatement potential, and Figure 1 visualises these results in standard MAC curve format with bar width proportional to abatement potential and bar height to MAC value. Negative MAC values indicate net economic benefit. Biogas recovery consistently emerges as the most cost-effective measure in biomass-rich sub-sectors (Hansupalak et al., 2016), while biomass cogeneration from bagasse delivers the largest abatement potential at the most favorable MAC in sugarcane processing (Li et al., 2024). The theoretical maximum abatement from all 15 measures at 100% adoption is 3,578 ktCO₂e. The Aligned Transition central estimate of 3,399 ktCO₂e is derived by applying a 95% feasibility weighting (3,578 × 0.95). The Monte Carlo simulation places this estimate within a 90% uncertainty interval of [2,780, 3,910] ktCO₂e and indicates that the rank order of the seven negative-MAC measures is preserved in 91% of simulation runs, with biomass bagasse cogeneration retaining rank-1 status in 99.4% of runs. Sensitivity analysis: ±20% CAPEX variation shifts individual MAC values by ±4–8 USD/tCO₂e; all negative-MAC measures remain cost-saving in the high-cost case.

The four most consequential parameter assumptions are (i) the social discount rate (8%), chosen to match the Thai Ministry of Finance benchmark for public-investment appraisal and consistent with Teagasc (2023) and USDA (2023) MAC studies; (ii) the technology lifetime (10–25 years per Table 3), drawn from BOI investment-promotion technology depreciation schedules; (iii) the IEA Thailand grid emission factor (0.4768 tCO₂/MWh, 2022); and (iv) the avoided-fuel reference prices used in the revenue-offset term (Appendix). One-at-a-time sensitivity testing on each yields the following: raising the discount rate to 12% (representative of SME commercial lending) shifts negative-MAC measures upward by 3–7 USD/tCO₂e, with sugarcane biomass cogeneration remaining negative at approximately −8.4 USD/tCO₂e but biogas-from-vinasse moving close to zero; raising it further to 15% pushes biogas-from-vinasse and rice-husk cogeneration into positive private-MAC territory. A 25% reduction in the grid emission factor (consistent with anticipated post-2030 EGAT decarbonisation) lowers the cold-chain electrification MAC by approximately 18 USD/tCO₂e but does not alter the rank order of negative-MAC measures. ±25% variation in the avoided-coal price band (USD 70–120/t) shifts the sugarcane biomass-cogeneration MAC by approximately ±3.5 USD/tCO₂e; under the lowest-coal-price case (USD 70/t) sugarcane biomass cogeneration retains a MAC of approximately −10.7 USD/tCO₂e. Shortening biogas digestor technology lifetimes from 15 to 10 years raises the cassava biogas-recovery MAC by approximately 6 USD/tCO₂e but the measure remains negative at approximately −2 USD/tCO₂e. The qualitative conclusion that biogas recovery and biomass cogeneration anchor the negative-MAC zone is therefore robust to all four parameter perturbations under social-discount-rate framing, but is sensitive to the discount-rate assumption when reframed against private SME financing costs.

4.3 BCG alignment index results

Table 5 presents the BCG-AI scores. No sub-sector achieves ‘BCG Advanced’ status (BCG-AI > 70), corroborating Jaroenkietkajorn et al.’s (2024) finding that persistent financing and technology barriers constrain BCG implementation. All five composite scores are arithmetically verified (e.g., Sugarcane: (74.2 + 65.8 + 65.2) / 3 = 68.4). Scores represent indicative rankings based on secondary proxy indicators and require empirical validation against primary facility data.

4.4 NDC gap analysis and scenario comparison

Table 6 presents the GHG reduction achievable under each scenario. A sector-level analytical proxy NDC target of approximately 2,910 ktCO₂e by 2030 is derived from Thailand’s 33.3% economy-wide reduction target. This figure is not an officially designated policy target: Thailand’s NDC and BCG documents do not publish a binding sectoral allocation for food processing, and the 2,910 ktCO₂e value represents a proportional allocation of the 184.8 MtCO₂eq economy-wide target by the food processing share of national industrial emissions. Proportional allocation has well-known limitations: it ignores differential abatement-cost curves across sectors, may misalign with cost-optimal allocation under a national carbon market, and does not reflect explicit equity considerations. The conclusion that the Aligned Transition scenario exceeds the proxy is therefore an analytical benchmark rather than a claim of formal policy compliance. The Aligned Transition central estimate of 3,399 ktCO₂e exceeds this proxy target and remains above 2,910 ktCO₂e in 84.3% of Monte Carlo simulation runs.

The 2,910 ktCO₂e value used as the sector-level NDC reference is, to re-emphasise, an analytical proxy rather than an official policy allocation; this caveat is now flagged in every section in which the figure appears. Thailand’s Office of Natural Resources and Environmental Policy and Planning (ONEP), the Department of Climate Change and Environment (DCCE), and the BCG strategy documents have not, as of the time of writing, published a binding sector-level GHG-reduction allocation for the food-processing industry. The proxy is computed by applying the food-processing share of national industrial emissions (approximately 1.575% of 184.8 MtCO₂eq) to the unconditional economy-wide 33.3% reduction target. This proportional method has three explicit known limitations: (a) it implicitly assumes uniform abatement-cost curves across all sectors, contradicting the very rationale for sector-specific MAC analysis; (b) it does not reflect the cost-optimal allocation that would arise under a hypothetical economy-wide carbon market or marginal-cost-equalised reduction scheme; and (c) it does not embed explicit equity weighting between large-emitter and smaller-emitter sectors. Statements in this manuscript that the Aligned Transition scenario “meets” or “exceeds” the proxy should therefore be interpreted as analytical benchmarks for cross-scenario comparison rather than as claims of formal policy compliance. If and when ONEP or DCCE publishes a formal sector-level NDC allocation for food processing, that figure may differ materially from the proxy used here, and the scenario findings would in that case need to be re-evaluated against the official allocation.

5 Discussion

5.1 Why biomass-rich sub-sectors outperform refrigeration-intensive ones

The MAC analysis reveals a structural divergence between biomass-rich sub-sectors (sugarcane and cassava), which dominate the negative-MAC zone, and refrigeration-intensive sub-sectors (frozen seafood), which populate the high-MAC tail. This divergence reflects three reinforcing factors. First, biomass-rich sub-sectors generate co-located residue streams bagasse, husk, and vinasse whose opportunity cost is near zero, so any energy-recovery measure displaces purchased fossil fuel at a one-to-one rate. Second, the capital intensity of biomass cogeneration is amortized over 20–25-year asset lives that align with mill replacement cycles, lowering the capital recovery factor burden per unit of abatement. Third, refrigeration-intensive sub-sectors face an electrification penalty because Thailand’s grid emission factor (0.4768 tCO₂/MWh) means electricity-substitution measures abate emissions only partially a structural ceiling that will not lift until grid decarbonization advances.

Sensitivity analysis indicates that a 25% grid emission factor reduction would lower the cold-chain electrification MAC by approximately 18 USD/tCO₂e, approaching the threshold for market-driven adoption (Tvaronavičienė, 2024). Meng et al. (2024) independently confirmed that changes in energy structure are a primary lever for decoupling agricultural carbon emissions from economic growth.

The implication for tropical export economies more broadly is that decarbonization sequencing should follow available residue streams first, while cold chain decarbonization should be timed to follow rather than preceding grid greening.

5.2 Comparative implications beyond Thailand

The MAC patterns observed in Thailand’s biomass-rich sub-sectors are broadly consistent with Teagasc (2023) for Irish dairy processing, where anaerobic digestion likewise occupies the negative-MAC zone, and with USDA (2023) for U. S. agriculture, where on-farm methane capture leads cost-effectiveness rankings. The framework presented here can therefore plausibly be extended to other ASEAN agri-food economies with similar residue-rich profiles Vietnam (cassava, rice), Indonesia (palm oil mill effluent, sugar), and the Philippines (sugar, coconut) although direct transfer of MAC values requires country-specific recalibration of grid emission factors, factor prices, and discount rates. The BCG Alignment Index, by contrast, is more Thailand-specific because the BCG framework itself is a Thai national construct, but its three-dimensional structure could be re-parameterized for analogous regional bio-circular-green policy frameworks elsewhere in the Global South.

The cross-country transferability claim warrants more careful qualification. While the methodological framework gate-to-gate baseline construction, eight standardised abatement measures, sector-specific MAC curves, Monte Carlo uncertainty bounding, and a composite Bio-Circular-Green index can in principle be re-applied to other ASEAN agri-food economies, the numerical MAC values reported here are not directly transferable. Country-specific recalibration is required for at least seven parameters whose magnitudes differ materially across the region: (i) grid emission factor (Vietnam approximately 0.62 tCO₂/MWh, Indonesia approximately 0.81 tCO₂/MWh, the Philippines approximately 0.57 tCO₂/MWh, all higher than Thailand’s 0.4768); (ii) commercial discount rates and SME lending costs; (iii) labour-cost components of OPEX; (iv) avoided-fuel reference prices, particularly coal and fuel oil; (v) residue-availability ratios, which depend on local cropping patterns and competing residue uses (e.g., bagasse for paper-pulp in some markets); (vi) policy-incentive coverage equivalent to Thailand’s BOI Section 31/33 framework; and (vii) regulatory frameworks for biogas-to-grid feed-in and PPA terms. Furthermore, the BCG framework itself is a Thailand-specific national policy construct; analogous regional frameworks (Indonesia’s Blue-Economy roadmap, Vietnam’s National Green Growth Strategy 2021–2030, the Philippines’ Sustainable Consumption and Production Action Plan) share thematic overlap but differ in policy architecture and implementation instruments. Transferability is therefore framed as methodological rather than numerical: the framework is portable, but country-specific numerical results must be re-derived locally on the basis of national audit data rather than transplanted from this Thai analysis.

5.3 Implementation barriers despite favorable MAC values

Several persistent barriers explain why negative-MAC technologies remain under-deployed in Thai practice despite favorable economics, consistent with Jaroenkietkajorn et al. (2024) and Mudzielwana (2025). Financial barriers are the most acute: SME processors face commercial lending rates of 12–18% that effectively raise their private discount rate well above the 8% social discount rate used in this study, shifting many measures from negative to positive private MAC. Technology and skills barriers also constrain deployment, as biogas systems require specialized operations and maintenance skills that are unevenly distributed across Thai provinces.

Institutional barriers arise where mills lease land or receive sugarcane on tolling contracts, weakening the alignment of capex recovery incentives with the operating entity. Regulatory barriers include opaque power purchase agreement terms for biogas-to-grid feed-in and limited monetization of negative-MAC measures under Thailand’s T-VER voluntary emission reduction scheme. These barriers explain the gap between technical and economic potential and motivate the policy recommendations developed below.

The “negative-MAC” label used in this study reflects social cost–benefit at an 8% public-sector discount rate rather than the private financial calculus actually faced by Thai food processors. Five concrete adoption barriers, drawn from Jaroenkietkajorn et al. (2024) and Mudzielwana (2025) and consistent with the implementation experience of BOI energy-efficiency programmes 2015–2022, explain why these technologies remain under-deployed despite favourable social economics: (i) commercial lending rates of 12–18% in the Thai SME segment raise the effective private discount rate well above 8%, shifting many measures from negative to positive private MAC; (ii) split-incentive problems where mills receive cane on tolling contracts and do not capture downstream cogeneration revenue; (iii) opaque feed-in-tariff and power-purchase-agreement terms for biogas-to-grid sales, with average T-VER project registration timelines of 18–24 months that erode net present value; (iv) under-developed third-party operations-and-maintenance markets for biogas digestors outside the Central region, raising perceived technology risk in remote facilities; and (v) limited managerial awareness of carbon-pricing instruments and CBAM-readiness benefits in SME processors. The headline claims of this manuscript are therefore explicitly softened: the negative-MAC findings should be read as “cost-effective in social terms” and “unlocked under the right financing and policy conditions” rather than as predictions of spontaneous market-driven adoption. Closing the gap between social and private cost-effectiveness is itself a primary purpose of the BOI investment-incentive and T-VER aggregation recommendations developed below.

5.4 BCG policy implications and strategic sequencing

The BCG-AI reveals that all five sub-sectors remain below the ‘BCG Advanced’ threshold, corroborating Jaroenkietkajorn et al.’s (2024) assessment. Policy recommendations include: (1) targeted BOI green investment incentives for biogas recovery in cassava starch and sugarcane mills; (2) mandatory carbon intensity reporting for facilities above 50 employees; (3) incorporation of the BCG-AI as a monitoring metric in the BCG Model’s 2026 review cycle; and (4) a sector-specific carbon credit aggregation mechanism under T-VER designed to allow SME processors to monetize negative-MAC measures whose individual scale is too small for stand-alone certification.

Strategically, negative-MAC measures should anchor the immediate (2026–2028) BCG implementation phase, while positive-MAC and grid-dependent measures should be sequenced into the 2028–2035 phase as the EGAT power mix decarbonizes. Mudzielwana (2025) recommended that overcoming financial constraints and policy fragmentation requires integrated climate-smart food system frameworks that align investment incentives with GHG reduction targets at the sub-sector level, a design principle directly reflected in the MAC-BCG integrated approach presented here.

5.5 Implications for carbon competitiveness

While the EU CBAM’s initial product scope does not yet include food products, scope expansion is anticipated (Clora and Yu, 2022; Pollitt et al., 2020; Schubert et al., 2026). The carbon intensity benchmarks and MAC curve results presented here provide an illustrative quantitative reference that Thai food exporters could draw on for voluntary, preparatory carbon footprint reporting; they are not intended as compliance-grade inputs to any future EU CBAM verification process. Within the present sample, the cassava starch and frozen seafood sub-sectors would, in an illustrative scenario, be the most exposed to future tightening of international carbon governance because their carbon intensities (468 and 312 kgCO₂e/ton, respectively) are an order of magnitude above sugar (55) and milled rice (42).

As an indicative orders-of-magnitude calculation, applying a hypothetical CBAM-equivalent price band of EUR 80–120 per tCO₂e to the Scope 1 + 2 emissions intensity of cassava starch implies a per-tonne carbon liability of approximately EUR 37–56, equivalent to roughly 8–12% of the typical FOB export price of native cassava starch (approximately USD 480–520 per ton in 2024). Equivalent calculations yield a 1–3% liability for refined sugar and 0.5–1.2% for milled rice. These figures are bound by data limitations and product-mix assumptions and are presented here only to indicate the order of magnitude of potential exposure; a full product-level CBAM exposure analysis covering HS-code disaggregated EU export shares, free-allocation phase-out trajectories, and competitor-country comparisons is identified as a priority follow-on study. Pearson et al. (2023) argued that establishing trusted, multi-lateral data exchange of standardized GHG measurements between food supply chain actors is an essential prerequisite for effective decarbonization strategy and for the verifiable carbon accounting that CBAM may eventually require a data infrastructure gap that the present study’s secondary data inventory (Table 1) begins to address at the sector level.

Expanding the indicative CBAM exposure analysis to the product level under the EU CBAM Implementing Regulation 2023/1773 default-value framework, and applying a hypothetical price band of EUR 80–120/tCO₂e to product-level Scope 1 + 2 carbon footprints, produces the following illustrative liability profile: (a) Native cassava starch (HS 1108.14) at the facility-level range 130.9–572.3 kgCO₂e/ton yields EUR 10–69/ton, equivalent to 2–14% of typical 2024 FOB prices (USD 480–520/ton). (b) Refined cane sugar (HS 1701.99) at 55 kgCO₂e/ton yields EUR 4.4–6.6/ton, equivalent to 0.8–1.5% of typical FOB prices (USD 450–550/ton). (c) Milled long-grain rice (HS 1006.30) at 42 kgCO₂e/ton yields EUR 3.4–5.0/ton, equivalent to 0.6–1.2% of typical FOB prices (USD 540–620/ton). (d) Frozen seafood, shrimp and fish products (HS 0303–0306) at 312 kgCO₂e/ton yields EUR 25–37/ton, equivalent to 1–3% of average FOB prices (USD 3,500–9,000/ton, depending on species and processing). (e) Canned tropical fruit, predominantly pineapple (HS 2008.20) at 185 kgCO₂e/ton yields EUR 15–22/ton, equivalent to 1–2% of typical FOB prices (USD 1,100–1,400/ton). Three caveats apply uniformly: (i) the calculations are bounded by Scope 1 + 2 only full Scope 3 inclusion would substantially increase liabilities, particularly for seafood and tropical fruit; (ii) the figures are highly sensitive to the still-undetermined CBAM default values for food products (the European Commission has not yet published Tier-1 default factors for HS chapters 10, 11, 17, or 20), to free-allocation phase-out trajectories, and to product-mix and HS-code disaggregation assumptions within each line; and (iii) these calculations are presented as orders-of-magnitude exposure indicators rather than compliance-grade liability estimates. Cassava starch and frozen seafood emerge as the highest-exposure sub-sectors on a percentage-of-FOB-price basis, providing a clear policy-priority signal for early CBAM-readiness investment in those two value chains. A full HS-code disaggregated CBAM-exposure model that incorporates EU export-share, default-value uncertainty distributions, free-allocation phase-out trajectories, and competitor-country (e.g., Vietnam, Indonesia) comparative carbon-intensity benchmarks is identified as a high-priority follow-on study, consistent with Pollitt et al. (2020), Clora and Yu (2022), and Schubert et al. (2026).

5.6 Limitations

Several limitations warrant acknowledgement. First, this study relies exclusively on secondary data, with three specific consequences: (a) facility-level heterogeneity is masked by sector-representative values, as already discussed in Section 3.2 with reference to the wide cassava starch carbon intensity range; (b) DEDE energy audit coverage is not uniformly distributed across sub-sectors, with canned fruit and frozen seafood having fewer audited facilities; and (c) OIE (2022) production volume statistics are reported at sector level and cannot capture intra-sector size variation. Future studies should incorporate primary field data from a stratified sample of at least 30 facilities per sub-sector.

Second, while the Monte Carlo simulation bounds parameter uncertainty for CAPEX, OPEX, emission factors, and adoption rates, residual model-structure uncertainty is not quantified; the expert-based MAC curve is sensitive to CAPEX assumptions: ±20% variation shifts individual MAC values by ±4–8 USD/tCO₂e, though all negative-MAC measures remain cost-saving in the high-cost case. Third, Scope 3 emissions are excluded from the primary analysis; inclusion of upstream agricultural emissions would substantially alter relative carbon intensities (Nguyen et al., 2017; Moungsree et al., 2023).

Fourth, the BCG-AI is a first-generation, proof-of-concept tool whose sub-indicator proxy variables are derived from secondary data and have not been empirically validated against primary facility-level assessments. Fifth, the indicative CBAM exposure figures presented above are forward-looking illustrations rather than predictions and should not be interpreted as a substitute for product-level CBAM compliance modelling.

6 Conclusion

This study presents one of the first sector-wide MAC curve analyses for Thailand’s agri-food processing industry published in the international peer-reviewed literature. By constructing MAC curves for five major sub-sectors using exclusively secondary quantitative data, this study demonstrates that significant GHG abatement is achievable at negative or low positive cost through biogas recovery, biomass cogeneration, and motor efficiency improvements findings grounded in Trakulvichean et al. (2025), Hansupalak et al. (2016), Yuttitham et al. (2011), and Li et al. (2024). The key findings are fourfold. First, total Scope 1 + 2 GHG emissions from the five sub-sectors are estimated at 8,737 ktCO₂e per year (arithmetic verified). Second, the most cost-effective abatement pathway is biogas and biomass energy recovery, yielding MAC values of −14.2 to −8.2 USD/tCO₂e. Third, an analytical proxy for Thailand’s NDC 2030 sector-level target of approximately 2,910 ktCO₂e is achievable under the Aligned Transition scenario (central estimate: 3,399 ktCO₂e; Monte Carlo 90% uncertainty interval: 2,780–3,910 ktCO₂e; probability of meeting the proxy: 84.3%), but not under BAU or Moderate Transition pathways. Fourth, all sub-sectors remain in the ‘Transitioning’ or ‘Emerging’ BCG tier. The policy priority matrix is clear: targeted financial incentives for negative-MAC measures should be the immediate BCG implementation priority; mandatory carbon intensity benchmarking should be established; and the BCG-AI should be integrated into BCG monitoring frameworks pending empirical validation. Future research should extend this analysis to include primary facility-level data and apply the MAC-BCG framework to other Southeast Asian economies where food processing decarbonization evidence is similarly sparse (Reay et al., 2020; Pearson et al., 2023).

In summary, this study contributes to literature in three ways. First, it provides a harmonized GHG accounting framework spanning five Thai agri-food processing sub-sectors. Second, it operationalizes the integration between sector-level MAC analysis and Thailand’s BCG Economy Model. Third, it introduces a first-generation BCG Alignment Index that future research can refine through primary facility-level data, alternative weighting schemes, and expert elicitation.

References

• 1

  APEC (2022). Bangkok Goals on Bio-Circular-Green (BCG) Economy. 2022 Leaders’ Declaration. Singapore: Asia-Pacific Economic Cooperation.

  • Google Scholar

• 2

  BCG Thailand (2021). Bio-Circular-Green Economy Model: Background and Strategic Framework 2021–2026. Bangkok: Ministry of Higher Education, Science, Research and Innovation (MHESI).

  • Google Scholar

• 3

  BurkeD. T.HyndsP.PriyadarshiniA. (2023). Quantifying farm-to-fork greenhouse gas emissions for five dietary patterns across Europe and North America. Resour. Environ. Sustain.12:100108. doi: 10.1016/j.resenv.2023.100108

  • CrossRef
  • Google Scholar

• 4

  CCPI (2025). Thailand climate performance ranking 2026. Climate Change Performance Index. Available online at: https://ccpi.org/country/tha/

  • Google Scholar

• 5

  Climate Action Tracker (2025). Thailand country assessment. Available online at: https://climateactiontracker.org/countries/thailand/

  • Google Scholar

• 6

  CloraF.YuW. (2022). GHG emissions, trade balance, and carbon leakage: insights from modeling thirty-one European decarbonization pathways towards 2050. Energy Econ.113:106240. doi: 10.1016/j.eneco.2022.106240

  • CrossRef
  • Google Scholar

• 7

  CrippaM.SolazzoE.GuizzardiD.Monforti-FerrarioF.TubielloF. N.LeipA. (2021). Food systems are responsible for a third of global anthropogenic GHG emissions. Nat. Food2, 198–209. doi: 10.1038/s43016-021-00225-9,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 8

  DEDE (2022). Thailand Energy Efficiency Annual Report 2022. Bangkok: Department of Alternative Energy Development and Efficiency, Ministry of Energy.

  • Google Scholar

• 9

  FatimahY. A.GovindanK.SasongkoN. A.HasibuanZ. A. (2024). The critical success factors for sustainable resource management in circular economy. J. Clean. Prod.469:143084. doi: 10.1016/j.jclepro.2024.143084

  • CrossRef
  • Google Scholar

• 10

  FengS.LakshmananP.ZhangY.ZhangT.LiangT.ZhangW.et al. (2023). A comprehensive continental-scale analysis of carbon footprint of food production: comparing continents around the world. J. Clean. Prod.426:138939. doi: 10.1016/j.jclepro.2023.138939

  • CrossRef
  • Google Scholar

• 11

  FragkosP.FragkiadakisK. (2022). Analyzing the macro-economic and employment implications of ambitious mitigation pathways and carbon pricing. Front. Clim.4:785136. doi: 10.3389/fclim.2022.785136

  • CrossRef
  • Google Scholar

• 12

  HansupalakN.PiromkraipakP.TamthiratP.ManitsorasakA.SrirothK.TranT. (2016). Biogas reduces the carbon footprint of cassava starch: a comparative assessment with fuel oil. J. Clean. Prod.134, 539–546. doi: 10.1016/j.jclepro.2015.06.138

  • CrossRef
  • Google Scholar

• 13

  IEA (2023). Thailand Energy Policy Review 2023. Paris: International Energy Agency.

  • Google Scholar

• 14

  IPCC (2022). Climate Change 2022: Mitigation of Climate Change. Contribution of Working Group III to the Sixth Assessment Report of the IPCC. Cambridge: Cambridge University Press.

  • Google Scholar

• 15

  JaroenkietkajornU.GheewalaS. H.MungkungR.JakrawatanaN.SilalertruksaT.LecksiwilaiN.et al. (2024). Challenges and opportunities of bio-circular-green economy for agriculture. Circ. Econ. Sustain.4, 1729–1750. doi: 10.1007/s43615-024-00355-9

  • CrossRef
  • Google Scholar

• 16

  JonesM. W.PetersG. P.GasserT.AndrewR. M.SchwingshacklC.GütschowJ.et al. (2023). National contributions to climate change due to historical emissions of carbon dioxide, methane, and nitrous oxide since 1850. Sci. Data10:155. doi: 10.1038/s41597-023-02041-1,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 17

  LiK.ZhaoM.LiY.HeY.HanX.MaX.et al. (2024). Spatiotemporal trends of the carbon footprint of sugar production in China. Sustainable Prod. Consumption46, 502–511. doi: 10.1016/j.spc.2024.03.010

  • CrossRef
  • Google Scholar

• 18

  McKinsey and Company (2009). Pathways to a Low-Carbon Economy: Version 2 of the Global Greenhouse Gas Abatement Cost Curve. New York, NY: McKinsey and Company.

  • Google Scholar

• 19

  MengY.ShenY.WangW.LiuY.WangF.WangH. (2024). Who affects carbon emissions? Drivers and decoupling effects of agricultural carbon emissions: evidence from Sichuan, China. Front. Sustain. Food Syst.8:1441118. doi: 10.3389/fsufs.2024.1441118

  • CrossRef
  • Google Scholar

• 20

  MichelM.EldridgeA. L.HartmannC.KlassenP.IngramJ.MeijerG. W. (2024). Benefits and challenges of food processing in the context of food systems, value chains and sustainable development goals. Trends Food Sci. Technol.153:104703. doi: 10.1016/j.tifs.2024.104703

  • CrossRef
  • Google Scholar

• 21

  MoungsreeS.NeamhomT.PolprasertS.PatthanaissaranukoolW. (2023). Carbon footprint and life cycle costing of maize production in Thailand with temporal and geographical resolutions. Int. J. Life Cycle Assess.28, 891–906. doi: 10.1007/s11367-022-02021-4

  • CrossRef
  • Google Scholar

• 22

  MudzielwanaR. V. A. (2025). Climate-smart food systems: challenges and opportunities for mitigating greenhouse gas emissions and ensuring sustainable food production. Front. Sustain. Food Syst.9:1580516. doi: 10.3389/fsufs.2025.1580516

  • CrossRef
  • Google Scholar

• 23

  NguyenT. H.WilliamsS.PaustianK. (2017). Impact of ecosystem carbon stock change on greenhouse gas emissions and carbon payback periods of cassava-based ethanol in Vietnam. Biomass Bioenergy100, 126–137. doi: 10.1016/j.biombioe.2017.02.009

  • CrossRef
  • Google Scholar

• 24

  OECD (2024). A Growth-Friendly and Inclusive Green Transition Strategy for Thailand. OECD Economics Department Working Papers No. 1797. Paris: OECD Publishing.

  • Google Scholar

• 25

  OIE (2022). Industrial Indices and Annual Report 2022. Bangkok: Office of Industrial Economics, Ministry of Industry.

  • Google Scholar

• 26

  PearsonS.BrewerS.ManningL.BidautL.OnoufriouG.DurrantA.et al. (2023). Decarbonising our food systems: contextualising digitalisation for net zero. Front. Sustain. Food Syst.7:1094299. doi: 10.3389/fsufs.2023.1094299

  • CrossRef
  • Google Scholar

• 27

  PollittH.NeuhoffK.LinX. (2020). The impact of implementing a consumption charge on carbon-intensive materials in Europe. Clim. Pol.20:S74–S89. doi: 10.1080/14693062.2019.1605969

  • CrossRef
  • Google Scholar

• 28

  ReayD.WarnatzschE.CraigE.DawsonC.GeorgeD.NormanR.et al. (2020). From Farm to Fork: growing a Scottish Food System that doesn’t cost the Planet. Front. Sustain. Food Syst.4:72. doi: 10.3389/fsufs.2020.00072

  • CrossRef
  • Google Scholar

• 29

  RuizJ.Martin-MorenoJ. M.PerezR. (2023). Mid-term policy considerations of the EU green deal. Energ. Strat. Rev.50:101239. doi: 10.1016/j.esr.2023.101239

  • CrossRef
  • Google Scholar

• 30

  SandströmV.ValinH.KrisztinT.HavlíkP.HerreroM.KastnerT. (2018). The role of trade in the greenhouse gas footprints of EU diets. Glob. Food Secur.19, 48–55. doi: 10.1016/j.gfs.2018.08.007

  • CrossRef
  • Google Scholar

• 31

  SchubertK.PommeretA.RicciF. (2026). Confronting the carbon pricing gap: second-best climate policy. J. Assoc. Environ. Resour. Econom.13, 293–324. doi: 10.1086/739152

  • CrossRef
  • Google Scholar

• 32

  SenatlaM.MervenB.HughesA.CohenB. (2013). Marginal Abatement Cost Curves: Mitigation Decision Tool. Technical Guide. Cape Town: Energy Research Centre.

  • Google Scholar

• 33

  SfakianakiE. (2019). Critical success factors for sustainable construction: a literature review. Manag. Environ. Qual.30, 176–196. doi: 10.1108/meq-02-2018-0043

  • CrossRef
  • Google Scholar

• 34

  SmolM.MarcinekP.DudaJ.SzołdrowskaD. (2020). Importance of sustainable mineral resource management in implementing the circular economy (CE) model and the European green Deal strategy. Resources9:55. doi: 10.3390/resources9050055

  • CrossRef
  • Google Scholar

• 35

  Teagasc (2023). Marginal Abatement Cost Curve 2023: Executive Summary. Dublin: Teagasc Agriculture and Food Development Authority.

  • Google Scholar

• 36

  TGO (2023). Thailand GHG Inventory Report: Industrial Sector 2020–2022. Bangkok: Thailand Greenhouse Gas Management Organisation.

  • Google Scholar

• 37

  TrakulvicheanS.LerdlattapornR.LueangwattanapongK.ChokmongkolN.SaengchanK.SongkasiriW. (2025). Carbon footprint of cassava starch in Thailand: a comprehensive study. Clean. Responsib. Consumpt.18:100293. doi: 10.1016/j.clrc.2025.100293

  • CrossRef
  • Google Scholar

• 38

  TvaronavičienėM. (2024). The transition towards renewable energy: the challenge of sustainable resource management for a circular economy. Energies17:4242. doi: 10.3390/en17174242

  • CrossRef
  • Google Scholar

• 39

  UNDP Thailand (2023) in Global Boiling and Climate Change: GHG Emissions in Thailand, ed. UNDP (Bangkok).

  • Google Scholar

• 40

  UsubharatanaP.PhungrassamiH. (2015). Carbon footprint of cassava starch production in north-eastern Thailand. Procedia CIRP 29, 462–467. doi: 10.1016/j.procir.2015.02.031

  • CrossRef
  • Google Scholar

• 41

  USDA (2023). Marginal Abatement Cost Curves for Greenhouse Gas Mitigation on U.S. Farms and Ranches. New York, NY: Office of the Chief Economist, U.S. Department of Agriculture.

  • Google Scholar

• 42

  VandyckT.KeramidasK.SaveynB.KitousA.VrontisiZ. (2016). A global stocktake of the Paris pledges: implications for energy systems and economy. Glob. Environ. Change41, 46–63. doi: 10.1016/j.gloenvcha.2016.08.006

  • CrossRef
  • Google Scholar

• 43

  VargaJ.RoegerW.t VeldJ. (2022). E-QUEST: a multisector dynamic general equilibrium model with energy. Econ. Model.114:105911. doi: 10.1016/j.econmod.2022.105911

  • CrossRef
  • Google Scholar

• 44

  YodkhumS.SampattagulS.GheewalaS. H. (2017). Life cycle greenhouse gas evaluation of organic rice production in northern Thailand. Journal of Environmental Management157, 217–223. doi: 10.1016/j.jenvman.2017.03.004

  • CrossRef
  • Google Scholar

• 45

  YuttithamM.GheewalaS. H.ChidthaisongA. (2011). Carbon footprint of sugar produced from sugarcane in eastern Thailand. J. Clean. Prod.19, 2119–2127. doi: 10.1016/j.jclepro.2011.07.017

  • CrossRef
  • Google Scholar