An analytical framework for integrated scenario analysis — pro-growth, pro-ecological sustainability
David Johnson · CE Engine
23 May 2026 · Online (Google Meet)
GDP Growth & CO₂ Emissions by Transition Scenario — IEA / NGFS / IMF WEO (2020–2050)
Growth thesis: Analysis indicates 3–4 carbon-neutral or carbon-negative industries at AI-level scale are needed to sustain economic growth while achieving absolute emissions decoupling. Sources: McKinsey Global Institute; Morgan Stanley Space Report 2024; IEA NZE 2050.
Policy is made by governments, enforced through markets, and felt through ecosystems. Yet the models that inform it are built in isolation:
Stranded assets, mispriced carbon, misdirected capital — when models don't talk to each other, policy misprices risk and misallocates investment at a civilisational scale.
CE provides a unified analytical workspace that lets climate and economic models run together — not just side by side.
Scenario builder, model runner, and policy simulator — all in one browser-based platform. No installation, no code required.
Connects physical climate outputs (temperature, rainfall, sea level) to economic variables (GDP, inflation, fiscal capacity, asset prices).
Built for analysts, investors, and policy professionals who need quantitative answers — not academic papers.
CE is neither a green lobbying tool nor a fossil fuel apologist. It is a rigorous analytical framework built on three principles:
Economic growth is not the enemy of ecological sustainability. Productivity, innovation, and capital formation are the mechanisms of decarbonisation — not its obstacle.
Ecosystem services are real economic assets. Their degradation is an unpriced liability on every balance sheet. CE quantifies it.
Models are challenged, not venerated. Every output comes with confidence bounds, scenario comparisons, and explicit assumption disclosures.
Five integrated knowledge and analysis domains — each a standalone tool, all speaking the same modelling language.
Physical projections — temperature, precipitation, sea-level, extremes. CMIP6 / IPCC AR6.
Sector-by-sector climate exposure and transition risk — agriculture, energy, finance, and more.
Carbon pricing, regulation, fiscal tools — modelled as levers with quantified GDP and emissions outcomes.
Country-level energy mix, reserves, and transition pathways across 140+ countries.
How economists try to describe the world — and where they fall short
Every model above was built to answer economic questions — not ecological ones. Climate enters (if at all) as a static damage coefficient. Not a dynamic system.
Dynamic Stochastic General Equilibrium models are the dominant tool at central banks and finance ministries worldwide.
No mainstream DSGE model predicted the 2008 financial crisis. The IMF's own models showed the US and European economies as fundamentally sound in 2007. The models had no mechanism for a financial sector collapse.
IAMs like DICE, RICE, and PAGE attempt to price climate damage into economic models. They are the backbone of the Social Cost of Carbon.
William Nordhaus (Nobel 2018) argued that optimal warming was 3°C+ — because aggressive mitigation is more expensive than adaptation. This shaped US climate policy for two decades.
DICE's damage function estimates 2.1% GDP loss at 3°C. Critics (Stern, Weitzman, Burke et al.) argue this is catastrophically wrong — and that non-linear, irreversible damages are completely missing from the maths.
A damage function that maps temperature → GDP loss as a smooth, polynomial curve. No tipping points. No systemic risk. No ecosystem collapse.
How well have they actually performed at forecasting?
Despite accuracy limits, economic models provide structured reasoning, scenario comparison, and policy sensitivity analysis — which is exactly what CE builds on.
At a 5% discount rate, $1 of damage in 2100 is worth $0.007 today. The further out you push the catastrophe, the more you discount it away. Climate risk vanishes mathematically.
Temperature → GDP is modelled as a smooth polynomial. Real-world risk has thresholds, tipping points, and non-linear collapses that smooth curves cannot represent.
Ecosystem services — pollination, water filtration, carbon absorption, flood buffering — are not in GDP. Their loss is invisible to standard economic models until it causes a market event.
How scientists simulate the atmosphere, oceans, and land — and what they've gotten right (and wrong)
The 6th Coupled Model Intercomparison Project coordinates 49+ models from ~20 countries / 33+ institutions. Their outputs form the basis of IPCC AR6 (2021).
Shared Socioeconomic Pathways define plausible futures based on development and emissions trajectories:
Sustainability — rapid transition, strong global cooperation, renewables dominant by 2050. Warming: 1.5–2°C
Middle of the road — current trends continue, moderate mitigation. Warming: ~2.7°C
Regional rivalry — fragmentation, nationalism, slow transition. Warming: ~3.6°C
Heavy fossil fuel use through 2100. Warming: 4.4°C (central estimate). Used as a stress-test scenario — most analysts now consider it implausible as written but essential as a tail-risk bound.
Global Mean Surface Temperature Anomaly (°C vs 1951–80)
Observed: NASA GISS Surface Temperature | Model: CMIP6 multi-model mean
Most GCMs do not represent tipping element interactions — AMOC slowdown, Amazon dieback, permafrost feedback, ice sheet collapse. These are the scenarios that break economic models.
Global models operate at 50–100km grid cells. Extreme rainfall events, urban heat islands, and local flooding require <1km resolution — computationally impossible at global scale.
Climate models use fixed emissions pathways — they cannot model policy response, market reaction, or technological change. That is exactly the gap CE fills.
Climate models produce physical outputs (temperature, precipitation, sea level, drought index). But those outputs live in a world of assets, labour, trade, and fiscal capacity. Translating physics into economics is the unsolved problem — and CE's core contribution.
Why coupling climate and economic models is so hard — and why most attempts fail
Economic models run on quarterly cycles. Climate models run on decadal projections. Connecting them requires bridging fundamentally different temporal resolutions.
Climate outputs are in physical units (°C, mm/year, Wm⁻²). Economic outputs are in monetary units (GDP, debt/GDP, CPI). The conversion between them is the damage function — the most contested piece of maths in all of climate economics.
Climate models have physical uncertainty (ensemble spread). Economic models have structural uncertainty (wrong mechanisms). They compound.
Climate damages reduce GDP → reduced GDP reduces mitigation capacity → higher emissions → more warming → more damages. This is a positive feedback loop that most models ignore because it requires genuine two-way coupling.
CE implements transmission modules — explicit pathways from physical climate variables to economic sectors, asset classes, fiscal positions, and financial stability. Each pathway is auditable.
The CE transmission framework maps physical hazards to economic consequences across six pathways:
Heat stress reduces labour productivity. Drought cuts agricultural yield. Flooding destroys infrastructure. All reduce potential GDP.
Coastal property, fossil fuel reserves, and agriculture land face repricing as physical and transition risk materialises.
Disaster response, adaptation infrastructure, and revenue loss from climate-impacted sectors strain government balance sheets.
Stranded assets, credit losses on climate-exposed lending, and insurance market failures create systemic financial risk.
Climate disruption in one region propagates through global supply chains — semiconductor fab floods, food export bans, port closures.
Temperature rise increases cooling demand, reduces thermal plant efficiency, and changes hydropower availability — restructuring energy costs globally.
Damage(T) = α·T² where T is temperature rise above pre-industrial. This implies:
Nordhaus called this manageable. The world continued developing.
Burke, Hsiang & Miguel (2015): using actual historical data, the relationship between temperature and GDP growth is non-linear and much steeper. At +4°C: 23% GDP loss globally, with poor countries losing far more.
Martin Weitzman argued the fat tails of climate outcomes dominate any cost-benefit analysis. Even if the probability of 6°C is 1%, the potential damage is so catastrophic and irreversible that standard expected-value calculations break down.
CE's position: We implement multiple damage functions and let users see the policy implications of each — rather than hiding the model choice.
Earth system tipping points represent non-linear, self-amplifying transitions that standard models ignore:
Tipping threshold: ~1.5–2°C. Committed to 7m sea level rise over centuries. Irreversible on human timescales.
Atlantic circulation weakening could cause rapid cooling in Europe (-3 to -8°C), sea level rise on US East Coast, monsoon disruption.
30–40% deforestation + warming could trigger a self-sustaining savannification of the Amazon — releasing ~150–290 GtCO₂, accelerating all other tipping points.
Arctic permafrost holds ~1,500Gt C. Thawing releases CH₄ and CO₂ — a 0.1–0.3°C warming contribution by 2100 (up to 1.5°C on century+ timescales) in high-emission scenarios.
Boreal fires and dieback shift the world's second-largest terrestrial carbon sink into a source. Underrepresented in all major models.
Tipping points interact. AMOC weakening can trigger Amazon dieback. Permafrost thaw accelerates all others. No IAM models this cascade.
How CE Engine integrates physical and economic models into actionable insights
Each layer outputs variables that feed into the next. A drought scenario reduces agricultural productivity (physical → macro), which reduces rural tax revenue (macro → fiscal), which limits adaptation spending (fiscal → physical).
Every number in CE has a traceable pathway. You can follow any output back to its source assumption — unlike black-box IAMs.
Current policies only — warming ~2.7°C by 2100
NDCs fully implemented — warming ~2.1°C
SSP1-1.9 pathway — warming held near 1.5°C
Carbon prices by jurisdiction (USD/tCO₂, 2026)
Gold bar = IMF minimum required for 2°C. Red = no federal price. Sources: World Bank Carbon Pricing Dashboard 2026.
The IMF estimates $75–$100/tCO₂ globally by 2030 is the minimum to hit 2°C. Every major emitter except the EU is well below this — or has no price at all.
Decoupling is real but insufficient at current rates. CE models the required acceleration, the investment needed, and the trade-offs between consumption today and ecological stability tomorrow.
CE also quantifies the growth that comes from decarbonisation — avoided damages, new industries, reduced import dependency, lower energy costs in the long run.
Quantifying the 52 GtCO₂e gap, the technology abatement stack, policy levers, and the welfare cost of inaction — all in one integrated model
Emissions trajectory — current policy vs. IPCC C1 net-zero pathway (GtCO₂e/yr)
Sources: UNEP Emissions Gap Report 2024; IPCC AR6 WG3 SPM Figure SPM.4 / C1 median trajectory.
At current national pledges, the world reaches ~50 GtCO₂e/yr by 2030 — the IPCC C1 pathway requires ~34 Gt by 2030. The gap widens every year of delay.
Remaining carbon budget for 1.5°C: 250 GtCO₂ from 2025. At current emissions, exhausted in ~4.4 years. Every year of delay costs ~57 Gt of that budget.
Abatement potential by technology class (GtCO₂/yr, 2035 base estimate)
Sources: IEA NZE 2050; IPCC AR6 WG3 Chapter 6; CE Emerging Technology Library.
Green hydrogen, direct air capture, advanced nuclear, enhanced geothermal — high potential, still expensive. Deployment speed determines if we hit 1.5°C or 2°C.
GtCO₂ already committed by existing fossil fuel infrastructure over its remaining economic lifetime, even if no new fossil development ever begins. Compared against the remaining 1.5°C and 2°C carbon budgets (IPCC AR6 WG1). Source: IEA WEO 2022, Global Registry of Fossil Fuels, Carbon Brief.
Oil & gas upstream infrastructure alone commits ~220 Gt (IEA WEO 2022 scope), nearly equalling the entire 1.5°C budget. The total committed emissions consume 59% of the 2°C budget. These emissions are locked in regardless of future policy unless assets are retired ahead of end-of-life.
Even if every government agreed today to stop all new fossil projects, the coal plants, gas furnaces, and gasoline cars that already exist would continue burning for their entire working lives. That’s nearly 3× the entire carbon budget before we hit 1.5°C.
When bars tower above the budget lines, it means retiring existing fossil infrastructure early isn’t optional — it may be mathematically required just to stay within budget, even before accounting for any new emissions.
Maximum abatement potential from commercially deployed technologies — solar PV, wind, electric vehicles, energy efficiency, nuclear fission, heat pumps, and geothermal — at full deployment, compared against the net-zero abatement requirement. Source: IEA NZE 2023, IRENA 2023.
Solar, wind, EVs, heat pumps, nuclear & geothermal cover ~88% of the gap but leave a ~5.4 Gt/yr shortfall in 2050 — the portion no commercially available technology can close at any deployment speed.
When the bars approach the green line, deployment speed — not new invention — is the binding constraint. The question is "how fast?" not "what?"
Each modelled with a separate abatement trajectory. A 0.82× de-duplication factor accounts for cross-technology overlap (e.g., EVs and grid decarbonisation share efficiency gains).
Stacked abatement from 13 emerging technologies + required breakthrough gap (GtCO₂/yr, base scenario)
Sources: CE Emerging Technology Library; IPCC AR6 WG3 Ch.6; IEA NZE 2023. 0.85× de-duplication factor applied.
Even stacking every known emerging technology at base-case deployment, a required breakthrough gap remains. The 13-technology portfolio does not close the 52 Gt problem alone — not even close in 2030.
Under the base scenario, the emerging portfolio grows from near-zero in 2025 to ~30 Gt/yr by 2050 — covering a large share of the remaining gap once mature technologies (solar, wind, EVs) do their part.
Green hydrogen, perovskite solar, BECCS, DAC, advanced nuclear fusion, ocean iron fertilisation, green steel, SAF, enhanced weathering and more — each a thin slice now, potentially transformative by 2040–2060.
Projected additional abatement by policy lever (GtCO₂/yr by 2035)
Calibrated from: IMF (2019) carbon pricing study; IEA NZE 2050; EU CBAM impact assessment 2023; IPCC AR6 WG3 Ch. 13.
At $150/tCO₂, a universal carbon price delivers the single largest non-technology abatement gain — and funds the green transition through revenue recycling. IMF minimum for 2°C: $75–100/t by 2030.
Steel, cement, shipping, and aviation are where price signals alone are insufficient. Sector-specific mandates — clean hydrogen standards, zero-emission vehicle sales quotas — are required for full decarbonisation.
The CE Policy Simulator shows that a full policy mix can deliver 18+ Gt/yr additional abatement by 2035 — covering most of the required gap alongside the technology stack.
The SCC is the present value, in USD, of all damages caused by emitting one additional tonne of CO₂ today — integrated over 100 years. It answers: what is inaction actually worth?
Market-rate discounting. Future generations valued at ~12 cents on the dollar relative to present. Implies modest near-term carbon prices are adequate.
U.S. EPA 2023 central estimate — doubled from prior $51/t. Calibrated to post-pandemic empirical Ramsey rate and Rennert et al. (2022) comprehensive damage review.
Near-zero pure time preference — future lives valued almost equally to present. Standard in European policy analysis. Justifies aggressive near-term action even at high abatement cost.
The EU ETS (~$72/t) sits between Nordhaus and EPA. The gap between current prices and what welfare science recommends is the unpriced externality at the heart of the climate-finance problem. The discount rate choice is an ethical decision dressed in mathematical clothing.
57 GtCO₂e/yr current baseline. Technology stack can close the gap — but only with the right economic signals and policy mandates.
A carbon price of $75–190/t bridges the gap between current market prices and the theoretically efficient level. Policy mix covers 18+ Gt/yr by 2035.
Annual welfare cost of the 52 Gt gap: $2.7T–$22.9T/yr depending on discount rate. At any defensible rate, inaction is more expensive than action.
No individual model captures this. An emissions model without an SCC misses welfare. An SCC without a policy lever model misses feasibility. A technology model without fiscal capacity analysis misses real-world constraints. CE connects all of these — transparently, interactively, and auditably.
Honest accounting of what we don't know — and why it matters for policy
Challenge: Markets systematically underprice long-duration, low-probability, high-consequence risks. Carbon is underpriced. Biodiversity loss has no price. Stranded assets are on balance sheets at book value.
Challenge: Some technologies (solar, wind, EVs) have delivered. Others (CCS, hydrogen at scale, nuclear fusion) remain expensive and limited. CE distinguishes between proven and speculative solutions.
Challenge: Beyond certain thresholds, adaptation options don't exist or are unaffordable. You cannot adapt to a 6-meter sea level rise. You cannot adapt agriculture in a world of +4°C ENSO events. Limits to adaptation are real and CE models them.
Challenge: Most climate-vulnerable nations have the smallest fiscal capacity, highest debt costs, and least access to green capital. CE explicitly models the finance gap for emerging markets.
The choice of discount rate for climate policy is not a technical question — it is an ethical question about how much we value future generations.
CE runs scenarios across the full range of discount rate assumptions and shows users the policy implications of each choice. We do not pick one — we make the assumption explicit and visible.
r = δ + η·g
r = social discount rate, δ = pure time preference, η = inequality aversion, g = per-capita growth rate
Every parameter is a moral choice, not a fact.
CE treats uncertainty as a feature to be quantified, not a weakness to be hidden.
We know the model structure but not the exact values. Climate sensitivity: 2.5–4.0°C per CO₂ doubling. Economic damage coefficients: 0.5–5× depending on study.
CE response: Ensemble runs across parameter ranges
We might have the wrong model structure entirely — missing feedbacks, wrong functional forms, ignored mechanisms.
CE response: Multiple competing model frameworks, user-selectable
Things we don't know we don't know. Tipping point cascades. Geopolitical responses. Technology breakthroughs or failures. Social tipping points.
CE response: Scenario stress-testing and tail-risk framing
Every CE output displays: the model used, the key assumptions, the confidence range, and a link to the underlying data source. No black boxes.
Give practitioners, investors, and policy teams the tools to ask: "What happens to my portfolio / economy / country if this policy is implemented under these physical conditions?" — and get a rigorous, auditable answer.