Climate Scenarios (NGFS)

Get the path of an economic or financial variable in a climate scenario of the Network for Greening the Financial System (NGFS), the scenarios central banks and supervisors use for climate stress tests, yearly from 2022 to 2050 for around fifty countries and regions.

The NGFS scenarios combine an integrated assessment model, which projects the energy system, emissions and the carbon price (model=”REMIND-MAgPIE”, “GCAM” or “MESSAGEix-GLOBIOM”), with NiGEM, the macroeconometric model of the National Institute of Economic and Social Research, which translates each scenario into GDP, inflation, interest rates, equity prices, exchange rates and energy prices. NiGEM separates the transition risk (the policies and technologies of the move to a low-carbon economy), the physical risk (the damage of climate change) and both together (risk=”combined”).

The variables: “gdp” (2017 PPP US$ billion), “inflation”, “long_term_interest_rate”, “long_term_real_interest_rate”, “policy_rate”, “unemployment_rate”, “equity_prices” (index, 2017 = 100), “effective_exchange_rate” (index), “domestic_demand”, “exports”, “imports” (2017 PPP US$ billion), “oil_price” (US$ per barrel), “gas_price” and “coal_price” (US$ per barrel of oil equivalent), and “carbon_price”, which comes from the integrated assessment model itself: US$ of 2010 per tonne of CO2 for the world and the model’s regions, every five years to 2100.

By default the level is returned: the NiGEM baseline (a world without further climate change and policy, scenario=”Baseline”) with the scenario’s deviation applied. With deviation=True the deviation from the baseline itself is returned: relative for levels (-0.034 for 3.4% lower GDP) and in decimal points for rates (0.0094 for 0.94 percentage points higher). Rates are decimal fractions (0.041 for 4.1%).

The data is the Phase 5 vintage (November 2024), hosted by IIASA and retrieved without an API key; the NGFS licence allows commercial use with attribution but not the redistribution of substantial parts, so it is retrieved when needed rather than stored with the Finance Toolkit.

See definition: https://www.ngfs.net/ngfs-scenarios-portal/

Also known as: NGFS scenarios, climate stress test scenarios, transition risk, physical risk, carbon price path.

Climate Scenarios (NGFS) in Python

Install the Finance Toolkit with:

pip install financetoolkit -U

Then call get_climate_scenario as shown below.

from financetoolkit import Economics

economics = Economics(end_date="2025-12-31")

climate_scenario = economics.get_climate_scenario(
    variable='long_term_interest_rate',
    scenario='Net Zero 2050',
    countries=['United States', 'Germany', 'United Kingdom', 'Japan'],
)

climate_scenario.loc[['2025', '2030', '2035', '2040', '2045', '2050']]

Which returns:

Parameters

The function accepts the following parameters:

• variable (str, optional): The variable: “gdp”, “inflation”, “long_term_interest_rate”, “long_term_real_interest_rate”, “policy_rate”, “unemployment_rate”, “equity_prices”, “effective_exchange_rate”, “domestic_demand”, “exports”, “imports”, “oil_price”, “gas_price”, “coal_price” or “carbon_price” (units above). Defaults to “long_term_interest_rate”.
• scenario (str, optional): “Net Zero 2050”, “Below 2°C”, “Low demand”, “Delayed transition”, “Nationally Determined Contributions (NDCs)”, “Current Policies”, “Fragmented World” or “Baseline”. Defaults to “Net Zero 2050”.
• model (str, optional): The integrated assessment model behind the scenario: “REMIND-MAgPIE”, “GCAM” or “MESSAGEix-GLOBIOM”. Defaults to “REMIND-MAgPIE”.
• countries (list[str] | str | None, optional): The countries or regions to include. Defaults to None, which includes all of them.
• risk (str, optional): “combined”, “transition” or “physical”. Current Policies has physical risk only, which is also its combined risk. Defaults to “combined”.
• deviation (bool, optional): Whether to return the deviation from the baseline instead of the level. Defaults to False.
• rounding (int | None, optional): The number of decimals to round the results to. Defaults to None.

Related Environment