• Member States have proposed five amendments to MARPOL Annex VI ahead of ISWG-GHG 22 (1–4 September) to meet the 2023 IMO GHG Strategy, each building on the IMO NZF but with different approaches to targets, pricing, and flexibility mechanisms.

• Three of the five proposals deliver significant, broadly similar emissions reductions, converging on a similar range by 2050. A proposal from Liberia delivers relatively low emissions reduction beyond business as usual and the impact of the Japanese proposal largely depends on developments in the fuel market.

Not a subscriber yet? Sign up for Countdown newsletters here.

Proposals submitted by IMO Member States will form the basis for discussion at the upcoming 22nd Intersessional Working Group on GHG Emissions (ISWG-GHG 22). Understanding these proposals and their impacts is essential both for industry players deciding on future fleets and for IMO Member States deciding on how best to deliver the ambitions in the 2023 IMO GHG Strategy.

As you read through out analysis, keep in mind that all modeling has its limits. Results reflect a series of assumptions about the proposals themselves and the future global fleet. Changing the assumptions would change the results.

At MEPC 84, Member States were invited to work towards addressing concerns with the draft of the IMO Net-Zero Framework (IMO NZF). We assess five proposals: four Circular Letters (CL) as well as a proposal from Japan submitted for the upcoming Marine Environment Protection Committee (MEPC) meeting in December.

Proposed changes will still need to go through the adoption process, which could take place this year or next (Figure 1) (see our explainer on how the IMO process works).

Figure 1. Member States will consider amendments to the IMO Net-Zero Framework which were submitted as Circular Letters (CL) or as Marine Environment Protection Committee (MEPC) submissions.

The aim of this assessment is to add to the analytical picture of the Net-Zero Framework and its alternatives. The IMO's Comprehensive Impact Assessment (CIA), conducted by DNV, remains the most extensive study of policy scenarios. However, the CIA predates the five proposals, including the current draft IMO NZF. Our analysis adds to the CIA with analysis to support nuanced assessments of new proposals without providing policy prescriptions.

The Terms of Reference for the upcoming meeting notes that Member States will consider proposals "in line with the 2023 IMO GHG Strategy" (ISWG-GHG 22/1). We therefore draw on the principles in the Strategy to establish the dimensions for our assessment, shown in Table 1.

Table 1. Objectives outlined in the 2023 IMO GHG Strategy are used to define the dimensions of our assessment

To understand the impact, we use our integrated assessment model, Navigate. For the full methodology of our model including sensitivity ranges, see the Navigate page. As of July 2026, Navigate is open source, with code, inputs, and data available on GitHub.

We hold assumptions constant across proposals (Table 2) and then show results with sensitivity ranges that vary inputs such as biomass availability and infrastructure buildout speed.

One assumption worth highlighting is that our modeling includes only waste-based biofuels. The Center is currently expanding its work on ethanol as a marine fuel, and expects to publish findings on sustainability, availability, and cost from October, which will feed into future iterations of Navigate.

Table 2. Key assumptions for the modeling of scenarios. Further details on the model and assumptions can be found on our Navigate page.

We do not assess whether proposals contribute to a just and equitable transition, which will depend on the details of a fund or facility still to be finalized. Neither do we assess the practicality of proposals which may include provisions that pose distinct challenges for implementation.  

Figure 2. An illustration of our assumption of the Tier 2 Remedial Unit price as always equal to the Marginal Abatement Cost. As targets set increasingly strict targets they set

As shown in Table 3, all proposals retain the IMO NZF's fuel standard architecture. Three of the five proposals retain revenue for zero- and near-zero energy (ZNZ) and a just and equitable transition, while Liberia's proposal removes revenue generation. Japan proposes an approach to funding for projects without GHG pricing.

Table 3. A comparison of the proposals looking at key elements of the regulation.

Figure 3: GHG Fuel Intensity (GFI) trajectories under the three IMO NZF variants, showing Tier 1 (Direct) and Tier 2 (Base) compliance pathways from 2029 to 2050 against a reference intensity of 93.3 gCO2eq/MJ.

“NZF As-Is” - Australia, Canada, South Africa, and UK 

This proposal makes no substantive changes to the framework approved at MEPC 83. It removes the 2028 target to account for the one-year delay, starting with the 6% reduction target in 2029.  

“NZF Phase-In” - Brazil 

Brazil proposes a phased approach by removing Tier 1 pricing for the first two years and shifting the overall timeline by two years. This is paired with a stricter 2041 target of 70% compared to 65% 2040 target in the NZF As-Is, creating a back-loaded trajectory with shallower targets in the beginning which become steeper after 2035 to reach the 70% target (Figure 2). 

A second key difference in Brazil’s proposal is the option to change the IMO Net-Zero Fund to a “facility.” Brazil clarified in its ISWG-GHG 22 submission that the distinction is whether implementation sits within the IMO as a fund, or with external institutions operating it subject to IMO guidelines. Under this model, a separate governing board like the structure used for the International Oil Pollution Compensation Fund, could retain authority over strategy and operations, while an international financial institution, such as the World Bank, could host and administer the facility as trustee. Brazil also proposes a sunset clause for the facility and the option of voluntary donations. 

“NZF Enhanced Pricing” - Tuvalu  

This proposal keeps the IMO NZF's two-tier structure but changes the Tier 1 to a 100% reduction and raises the Remedial Unit 1 price from 100 to 300 USD/tCO2eq. This changes Tier 1 to a pure tax on emissions up to Tier 2. Surplus Unit trading and banking are removed, requiring ships to lower emissions or purchase Remedial Units.

Figure 4. The GFI under the Market-Defined GFS from 2029 to 2050. In Scenario 1 LNG is the only fuel meeting the viability criteria with a 13% GFI reduction by 2050. A lower-emissions fuel qualifies and re-anchors the minimum in Scenario 2, reaching a 37% reduction by 2050.

While other proposals use targets to incentivize the development of a market for alternative fuels, Liberia’s proposal defines targets based on the existing market. It replaces the two-tier structure with a single tier that permits Surplus Unit trading but removes Remedial Units as a compliance option.  

The targets are set by a 30-year trajectory from the fleet’s weighted-average fuel intensity to the intensity of the lowest-emissions “commercially viable” fuel. Viability is determined by availability, affordability, and scalability criteria, and is reviewed every five years. 

Ambiguity in the criteria makes it hard to project which fuels will qualify. Looking at current fuel markets, we find that LNG is likely the only fuel which would meet the criteria. To explore what would happen if the fuel market were to change, we model two illustrative GFI trajectories: 

• Scenario 1: LNG is the only fuel that meets all three criteria, setting the GFI minimum. 

• Scenario 2: LNG sets the GFI minimum from 2029, but we assume that at the first review in 2034, a lower-emissions fuel qualifies and re-anchors the trajectory over a new 30-year path to 2064. As an illustrative endpoint, we use a 70% reduction (28 gCO2eq/MJ), based on a low-emissions fuel threshold used in the EU for low-carbon and e-fuels (RED III, Article 29a). The trajectory reaches 37% reduction by 2050. 

An ISWG-GHG 22 submission from Liberia and Panama also proposes allowing ships to earn Surplus Units for verified GHG reductions from energy efficiency improvements. This would expand the pool of tradeable Surplus Units, likely lowering their price. For simplicity, we have not included this option in our modeling of the Liberia proposal.

Figure 5. The GFI under the Two-Tier Market-Defined proposal from 2029 to 2050. In Scenario 1 no low-emissions fuel meets the commercial viability criteria and targets flatline. In Scenario 2, a fuel that is 70% below baseline sets the 2050 target.

The Japanese proposal maintains two tiers from the IMO NZF while using the market-defined logic of the proposal from Liberia, attempting to merge priorities of Member States. It sets new, less stringent, targets based on assumed replacement rates and fuel consumption of the global fleet through 2035. It also removes the longer-term targets and instead states that beyond 2035, the targets will be set, “taking into account the global availability, affordability and scalability of new fuels and technologies” (MEPC 87/7/1, 35.3). 

As with the Liberia proposal, the difficulty of forecasting the fuel market makes defining a single set of post-2035 targets challenging. We therefore model two scenarios: 

• Scenario 1: No fuel meets the criteria beyond what's already in the 2035 fuel mix. Without a fixed target to create market pull for lower-emissions fuel, targets flatline through 2050.  

• Scenario 2: We assume a lower-emissions fuel qualifies and sets a new target by 2050. We use an illustrative 70% reduction to set the end point in 2050. 

Japan's submission also includes an alternative design where Tier 1 remains fixed at 21.9% through 2035, instead of the 13% gap between Tier 1 and Tier 2. This means Tier 1 contributions would be higher at first and lower over time, potentially phasing out in 2035. For simplicity, we model only the constant 13% gap. 

An important consideration for both Liberia and Japan’s proposals is whether a low-emissions fuel is expected to become commercially viable in the market without a guaranteed incentive from the IMO.

Figure 6: Modelled greenhouse gas emission reductions from 2029 to 2050 under the four proposals, shown as overlapping uncertainty bands against a shared set of modelling assumptions. For the two Market-Defined proposals, we show both scenarios.

Findings on emissions reductions: 

• The three NZF variants deliver broadly similar emissions reductions relative to the reference, converging on a range of roughly 60–85% reduction by 2050. 

• NZF Phase-In (Brazil) delivers slightly lower abatement through 2040, reflecting its delayed and shallower near-term targets, while the steeper post-2035 trajectory closes most of the gap by 2050. 

• NZF Enhanced Pricing (Tuvalu) achieves the deepest near-term cuts due to the pure tax. Our modelling did not include revenue recycling, which we would expect to drive additional reductions.  

• In the Market-Defined GFS (Japan), emissions from Scenario 1 are largely aligned with business as usual. Scenario 2 leads to a 13–20% reduction by 2050. 

• There is a sharp divergence in Japan’s Two-Tier Market Defined proposal between Scenario 1, which achieves limited reductions, while Scenario 2 delivers reductions between 60–70% in 2050. As noted, the viability of scenario 2 requires fuel with 70% emissions savings to be widely available in the 2030’s.

Any modelling of the future fuel mix comes with a high degree of uncertainty. Our aim is not to predict the future fuel mix, but to explore how a range of low-emission fuels could enter the market given our set of assumptions. We group fuel types both for simplicity and because they supply similar markets (i.e., blue and green ammonia) or rely on similar feedstocks (i.e., biofuels, e-fuels with carbon). 

Figure 7: Modelled global fuel consumption between 2026 and 2050 (median case) across the three IMO NZF variants.

Findings on the fuel mix for NZF variants: 

• Fuel-mix outcomes are broadly similar across the three NZF variants. Ammonia and biofuels emerge as major low-emissions fuels, although this should be read as indicative according to our assumptions.  

• In 2050 all three NZF variants show significant shares of non-fossil fuel, representing roughly 80 - 85% of the fuel mix in the median case.  

• We also see higher uptake of energy efficiency under Tuvalu's Enhanced Pricing as a result of the pure tax, keeping energy demand roughly constant despite assumed trade growth of 1–2.5% CAGR across segments – in line with the CIA's low growth scenario (Table B-2 in the Fourth IMO GHG Study). 

• The three NZF variants include GHG pricing to support a just and equitable transition and rewards for ZNZ fuels – none of which we model here. Because incentives like these can strengthen the business case for ZNZ fuels, including them would likely shift the fuel mix further toward lower-emissions options than shown above.

Figure 8: Modelled global fuel consumption between 2026 and 2050 (median case) for business as usual, and the two scenarios of the Market-Defined GFS.

Liberia’s Market-Defined GFS, in contrast, is notably similar to BAU (Figure 6). Scenario 2 is also dominated by fossil fuels, with ammonia representing 25% of the mix in 2050.

Figure 9: Modelled global fuel consumption between 2026 and 2050 (median case) for BAU, and the two scenarios of the Two-Tier Market Defined.

In Scenario 1 of Japan’s proposal, where the targets flatline, the share of lower-emissions fuels decreases over time because of an assumption that ammonia emission factors improve over time, lowering the need for alternative fuel. Alternatively in Scenario 2, the fuel mix reflects the 70% emissions reduction target in 2050.  

Transport costs include ship capital and operating expenses, plus two types of compliance cost: the additional cost of switching to lower-emissions fuels, and compliance payments. Figure 10 shows fuel-switching costs and Tier 1 payments, which are designed to stay below the cost of switching fuels. It excludes cost impacts and revenue recycling from Tier 2 payments, which depend on fuel availability, and the trajectory beyond 2040. 

Figure 10: Increase in transport cost relative to the BAU scenario and the GHG reductions in 2040.

Findings on the transport costs and GHG reduction in 2040: 

• With higher reductions comes higher transport costs, largely due to the additional cost of fuel switching. 

• Across the NZF variants there is a similar range of emissions reductions but higher potential costs for Enhanced Pricing. However, we did not include revenue recycling, which could contribute to lower transport costs and higher emissions reductions. 

• The Japanese proposal shows a high range of both costs and GHG reductions, reflecting the range of possible outcomes in their proposal. 

Limitations and further considerations on transport costs: 

• Our modeling does not capture revenue recycling. The NZF variants collect revenue intended for reuse, as rewards for zero- and near-zero fuels and to support a just and equitable transition for Member States.  

• These figures reflect only the direct cost of moving goods by sea. They do not capture economy-wide effects from either the added costs or the recycled revenue. These effects have been shown to vary across countries by the types of goods traded, transport distances, and the quality of freight infrastructure.

For a more detailed treatment of transport cost impacts, see Scenario 24 in the IMO’s Comprehensive Impact Assessment.

Table 4. A comparison of the proposals across the dimensions mapped onto the 2023 IMO GHG Strategy.

Findings on emissions reduction: 

• Outcomes are broadly similar across the three NZF variants on emissions and fuel mix but diverge on cost.  

• Japan’s Two-Tier Market-Defined proposal also achieves significant reductions and high uptake of non-fossil fuels in Scenario 2. This depends on the assumption that a fuel with 70% emissions reduction meets the criteria and sets the 2050 target.  

• Liberia's proposal is constrained by the 30-year horizon. Even if targets are re-anchored, it produces a shallower trajectory that delivers weaker emissions outcomes, while also not generating funds to recycle for ZNZ fuels and a just and equitable transition. 

• We also show that transport costs increase when revenue from Remedial Unit 2 deficits is factored in. These estimates are sensitive to assumptions on fuel availability, Remedial Unit pricing, and the trajectory beyond 2040 and do not include any expected impacts from recycling of revenue.

A key limitation in our modeling is the lack of revenue recycling. Tuvalu's Enhanced Pricing generates significant revenue generation which can be deployed for ZNZ fuels and a just and equitable transition. Revenue collection is also significant across the other NZF variants. Even partial recycling could unlock greater emissions reductions through ZNZ fuel support and reduce net transport cost impacts. 

As IMO Member States evaluate each of these proposals against the ambitions of the 2023 IMO GHG Strategy, these results do not provide the full picture, but they offer a useful understanding of the potential impacts that can be evaluated against ambitions in the 2023 IMO GHG Strategy. 

• Incentivizing low-emissions fuels: policy options and trade-offs 

• Mapping the adoption process for new drop-in fuels

• Biomethane supply potential and its relevance for shipping by 2030

• Ammonia Process Safety Management Maturity Self-Assessment Tool

• Ammonia Competency and Training Practitioner Handbooks