CIRCEE is a dynamic general equilibrium model, augmented with material flow analysis and mass balance, with a one-way soft-link to the IAM WITCH model (Emmerling et al., 2016) for energy efficiency, energy prices, emission factors and energy investments trajectories. It analyses the circular economy by integrating material, energy, and economic flows. The model simulates interactions among producers, consumers, government, and the external sector, with a particular focus on consumer-focused circular strategies (sharing, repair, sufficiency), policies (e.g., EPR schemes) and climate change mitigation goals, capturing the trade-offs and synergies between different sustainability objectives.
This repository also includes CIRCEE-LIFE, a two-way iterative coupling of CIRCEE with LIFE, a lifestyle-dynamics submodel (including cognitions, behaviours, and material and social context) that distinguishes three lifestyle groups from the typology of Pettifor et al. (2023). The coupling jointly calibrates two behavioural modifiers (sharing and sufficiency / lowering expenditures) at year 2020, then iterates between CIRCEE and LIFE until the lifestyle frequency trajectories stabilise out to 2060. The sufficiency and sharing lifestyles are included here; repair is available on demand and will be released open-source by the end of the year.
⚠️ Expect long runtimes and high memory use if you do not use an HPC.Version note. This repository holds the current version of CIRCEE. The 2025 Resources, Conservation and Recycling and 2026 Global Environmental Change papers were run on earlier versions: the CIRCEE calibration has since been refined as newer material- and waste-flow data became available, and the LIFE calibration was updated by Hazel Pettifor and Maureen Agnew. Those papers cannot be reproduced exactly from this code — the qualitative dynamics are unchanged, but levels can differ slightly. Earlier versions for the reproduction of the two papers will be made available soon.
# 1. Open config.sh and edit the scenario lines:
# RUN_MODE="coupled" # or "baseline" for no lifestyle heterogeneity
# SCENARIO_SHARING="ecoactive" # or "affordability"
# SCENARIO_SUFFICIENCY="ecoactive" # or "affordability"
# Then set MATLAB_BIN to point at your MATLAB binary.
# 2. Run:
bash run.shThe script runs a smoke test on MATLAB, then either runs the full coupling or a single baseline CIRCEE run, depending on RUN_MODE.
To run CIRCEE alone with the exogenous paths in data/<REGION>/raw/shocks.csv, from MATLAB:
cd src
CIRCEE_RunFileCIRCEE_RunFile.m writes src/CIRCEE_shocks.m before invoking Dynare, so calling dynare CIRCEE_PF.mod directly will fail on that missing include.
Reproducing Beyond Ownership: Lifestyles and inequality shape product-sharing's sustainability potential
The paper reports twelve runs: four lifestyle configurations (sharing driver × sufficiency driver) under three ecosystem scenarios, plus a zero-modifier reference run.
Environment. Results were produced with MATLAB R2024b and Dynare 6.5 on Linux (HPC). Other Dynare versions have not been tested; the perfect-foresight solver's behaviour can differ between major releases.
Producing a scenario. Set SCENARIO_SHARING, SCENARIO_SUFFICIENCY and SIGMA_SCENARIO in scripts/config.sh, then cd scripts && bash run.sh. Each configuration is a separate job; expect 2–3 days on an HPC. The coupling converges within two outer iterations for all twelve configurations.
Welfare. Once all runs are complete, edit the USER PATHS block in src/CIRCEE_WelfarePostProcess_batch.m and run it. This produces the consumption-equivalent variation, access ratios and distributional indices used in Figures 4 and 5.
Figures. See post/README.md.
A note on the converged modifiers. The behavioural modifiers in shocks_final_*.csv are reached iteratively by the coupling, each solve starting from the previous one. Supplying a converged file to a single classic-mode run asks the solver to reach the same allocation from the steady state in one step; this does not converge for the configurations with the largest behavioural response. Reproduce coupled runs through scripts/run.sh, not by feeding converged modifiers to classic mode.
On convergence warnings. run.sh reports "Modifier run did not converge"
when Dynare's perfect-foresight solver stops on its iteration or step-size
criterion rather than reaching its error tolerance. This concerns a single
CIRCEE solve, not the coupling: the CIRCEE–LIFE fixed point is judged by the
frequency changes in convergence_*.csv, and in the runs behind this paper
all twelve configurations converged there within two outer iterations. Where
the warning appeared — four of the twelve configuration — it was on the final iteration's solve or on the redundant re-solve that follows it
If you see the warning, check whether the solution actually satisfies the model before discarding the run. Adding resid; after the solver in CIRCEE_PF.mod reports the residual of every equation by name; all should be near zero. A failed solve leaves the solver's initial
guess in place: in a single classic-mode run that is the steady state, so the
output is then identical to the zero-modifier reference; inside the coupled
loop each run starts from the previous iteration's solution (which has not failed), so a failure
there produces a path that still differs from the reference.
Note that scripts/lib/Full_coupling.sh does not copy the final outputs when
this warning fires on the last run, so Sharing_final_*.csv and
Lowering_Expenditures_final_*.csv may be absent while the other outputs
are present.
On the joint calibration warning. Phase 1 tunes each behavioural modifier
in turn, holding the others at their current values, and each tuner converges
to its own target to within 0.01%. After each pass, check_joint_verification
reports how far the joint run — all modifiers active — sits from a target
computed from the zero-modifier baseline. Those two targets are not the same
quantity: the tuners work from a baseline in which the other modifiers are
already calibrated, the joint check from one in which none are. The check
therefore reports a gap in every run, including the twelve behind this paper,
and prints "Joint calibration did not converge. Proceeding with best available
modifiers."
This affects the diagnostic, not the calibration. Phase 2 runs on the tuner modifiers, which are the ones that reproduce the LIFE propensities. Computing the tuner targets from the zero-modifier baseline instead was tested and rejected: for the lowest-income group it implies a change of around 40% at T0, which the model does not solve.
The modifiers should therefore be read as reproducing the LIFE propensities conditionally — each with the others held at their calibrated values — rather than simultaneously.
Set RUN_MODE in config.sh (or override on the command line: RUN_MODE=baseline bash run.sh).
| Mode | What happens | Typical use |
|---|---|---|
coupled |
Full CIRCEE ↔ LIFE iterative coupling. Phase 1 jointly calibrates T0 modifiers in propensity space; Phase 2 iterates the LIFE dynamics until lifestyle frequencies converge. Hours-scale runtime. | The full model run. |
baseline |
Single CIRCEE run with all behavioural modifiers held at zero. No Phase 1, no Phase 2, no LIFE feedback. Minutes-scale runtime. SCENARIO_* settings are ignored. |
Sanity checks, sensitivity analysis on CIRCEE parameters, generating a reference trajectory. |
To run CIRCEE on its own, without any lifestyle scenario, set RUN_MODE="baseline" (or RUN_MODE=baseline bash run.sh). This runs the economic model once with all behavioural modifiers at zero and Current enablement infrastructures (Baseline in the model), producing the reference trajectory against which the coupled runs are compared. It is the fastest way to run the model and the recommended starting point.
Two behavioural narratives drive sufficiency and sharing lifestyle adoption. They can be combined freely across sharing and sufficiency:
| Scenario | Logic |
|---|---|
ecoactive |
Awareness-led. Adoption is highest among households that already exhibit environmental awareness and strong low-carbon cognitions. |
affordability |
Necessity-led. Adoption is driven by the need to save money and by budget constraints. |
All user-facing options are set in config.sh:
| Variable | Options | Meaning |
|---|---|---|
RUN_MODE |
coupled | baseline |
Full CIRCEE↔LIFE coupling, or a single CIRCEE run with no modifiers. |
SCENARIO_SHARING |
ecoactive | affordability |
Behavioural narrative for sharing (coupled mode). |
SCENARIO_SUFFICIENCY |
ecoactive | affordability |
Behavioural narrative for sufficiency (coupled mode). |
SIGMA_SCENARIO |
Baseline | Progressive | Regressive |
Infrastructure (σ^es) scenario — see below. Run Baseline first (welfare reference). |
FORESIGHT_MODE |
anticipation_errors | perfect_foresight |
How agents anticipate the shock path. |
WELFARE_MODE |
single | all |
Run welfare for the current σ scenario only, or all σ scenarios sequentially. |
SSP_SCENARIO |
SSP2 |
Socioeconomic pathway (SSP2 available now). |
MATLAB_BIN |
path | Path to your MATLAB binary. |
This yields four combinations, e.g. ecoactive_ecoactive, affordability_affordability, ecoactive_affordability, affordability_ecoactive. Output files are tagged with the combined name.
Independently of the behavioural narratives, the surrounding infrastructures is represented by the elasticity of substitution (siggma_es) between ownership-based and PSS-based (market) energy services. A higher siggma_es means home and PSS energy services are closer substitutes, i.e. more enabling infrastructures for moving away from ownership. They differ in the level of σ and in how it is distributed across the three lifestyle groups over 2018–2050:
| Infrastructure configurations | Logic |
|---|---|
Baseline |
Low substitution held roughly constant — a weakly enabling infrastructures. |
Progressive |
siggma_es rises over time, with the largest increase for lower-income groups (enablement concentrated on those who most depend on external conditions). |
Regressive |
siggma_es rises over time, with the largest increase for higher-income groups. |
Progressive (Broad in the paper) and Regressive (Concentrated in the paper) reach a similar aggregate level of enablement by 2050 but distribute it oppositely, isolating the effect of how enablement is distributed from how much there is. The σ trajectories are set in the calibration/shock inputs; see config.sh and the infrastructures-configuration block in the calibration files. (Earlier model versions also included Utopia_Equality and Dystopian_Equality.)
Two dimensions vary across runs — the ecosystem scenario (the substitution elasticity path) and the behavioural modifiers (the lifestyle configuration).
Output filenames encode both, and Baseline refers to the ecosystem, not to the absence of behaviour.
| File | Ecosystem | Behavioural modifiers |
|---|---|---|
NoModifiers_AE.csv |
Current enablement | all zero — the reference run |
<lifestyle>_AE_Baseline.csv |
Current enablement | as set by the lifestyle configuration |
<lifestyle>_AE_Progressive.csv |
Broad | as set by the lifestyle configuration |
<lifestyle>_AE_Regressive.csv |
Concentrated | as set by the lifestyle configuration |
AE denotes anticipation errors. <lifestyle> is one of ecoactive_ecoactive, affordability_affordability, ecoactive_affordability, affordability_ecoactive, given as sharing driver followed by sufficiency driver.
All results in the paper are reported as changes relative to NoModifiers_AE.csv: the Current enablement ecosystem with no lifestyle heterogeneity.
Code tokens differ from the labels used in the paper:
| Code token | Paper label |
|---|---|
Baseline |
Current enablement |
Progressive |
Broad |
Regressive |
Concentrated |
NoModifiers |
reference run |
lowcarbon |
Higher income |
cautious |
Medium income |
constrained |
Lower income |
- MATLAB R2024b (older versions probably work; R2023b or later is recommended).
- Dynare 6.5, installed and on MATLAB's path. Download from https://www.dynare.org/. The published results were produced with 6.5 on Linux (MATLAB R2024b); other versions have not been tested and the perfect-foresight solver's behaviour can differ between major releases.
- bash ≥ 4, plus
bc,awk,sed,grep(standard on macOS/Linux). - For distributional/figure post-processing: R (≥ 4.0) with
ggplot2,patchwork,ggh4x.
.
├── README.md
├── LICENSE
├── CITATION.cff
├── VARIABLES.md
├── .gitignore
├── post/ ← Figures
├── README.md
├── Figure_2.R
├── Figure_3.R
├── Figure_4.R
├── Figure_5.R
├── Supplementary_Figure_1.R
├── Supplementary_Figure_2.R
├── Supplementary_Figure_3.R
├── Supplementary_Figure_4.R
├── Supplementary_Figure_5.R
├── Supplementary_Figure_6.R
├── Supplementary_Figure_9.R
├── Supplementary_Figure_10-13.R
├── Supplementary_Tables_1_2.md
├── scripts/
│ ├── config.sh ← edit this (scenarios + MATLAB path + tolerances)
│ ├── run.sh ← run this
│ └── lib/
│ ├── parameters.sh ← LIFE model constants (β, α, frequency arrays, ratios)
│ ├── common.sh ← shared helpers (MATLAB call, shocks builder, output reader)
│ ├── Full_coupling.sh ← Phase 1 (T0 calibration) and Phase 2 (outer loop)
│ ├── tune_sharing.sh ← 1D bisection tuner for sharing modifiers
│ └── tune_expenditures.sh ← 1D bisection tuner for sufficiency modifiers
├── src/ ← model code (MATLAB + Dynare)
│ ├── MODEL_selection.inc ← region selection (@#define REGION)
│ ├── CIRCEE_PF.mod ← Dynare model
│ ├── CIRCEE_RunFile.m ← top-level MATLAB driver
│ ├── CIRCEE_steadystatemodel.m ← steady-state computation
│ ├── CIRCEE_calibration.m ← calibration values (generated/provided)
│ ├── CIRCEE_WelfareAnalysis.m ← CEV + distributional indices
│ ├── CIRCEE_WelfarePostProcess.m ← welfare driver (single run)
│ ├── CIRCEE_WelfarePostProcess_batch.m ← example driver (loops over runs; edit USER PATHS)
│ ├── CIRCEE_WelfareExport.m ← welfare/distribution CSV export
│ ├── Footprints.m ← carbon/waste/material footprints
│ └── clean_up.m ← housekeeping helper
└── data/
└── <REGION>/ ← e.g. JPN
├── raw/
│ ├── calibration.csv ← CIRCEE calibration table
│ └── shocks.csv ← exogenous paths; zero modifiers (classic mode)
└── templates/
└── shocks.csv ← skeleton for the coupled pipeline
The paths above follow scripts/config.sh (SRC_DIR=../src, PATH_TEMPLATES=../data//templates/, RESULTS_DIR=../results), so the scripts are run from within scripts/. Adjust config.sh if your layout differs.
results/ is created at runtime and is not tracked. Runtime-generated files — src/CIRCEE_shocks.m, src/CIRCEE_endvalues.m, src/CIRCEE_baseyear_values.m, src/data_shocks.csv, Dynare's src/+CIRCEE_PF/, and the entire results tree including welfare .mat files — are gitignored.
The two behavioural modifier triplets (sharing × {lowcarbon, cautious, constrained} and sufficiency × the same three groups) are jointly calibrated so that CIRCEE's year-2020 outputs match propensity-space targets in the LIFE model. Each block is tuned by 1D bisection (lib/tune_sharing.sh, lib/tune_expenditures.sh); the two blocks are then alternated until both are jointly satisfied. The order of tuning alternates each iteration to reduce ordering bias.
Propensities in LIFE are homogenous across activities: it is a property of the driver, not the behaviour. It says how ecoactive-driven or affordability-driven a lifestyle group is.
- Run A — CIRCEE with all behavioural modifiers = 0 (the homogeneous baseline, where all households behave identically). Runs once.
- Run B — CIRCEE with the full behavioural modifier path. Runs each iteration.
- LIFE update — for each year T on the 2025…2060 grid:
out_freq_h(T) = in_freq_h(T) × ES_h(T) / ES_h(T-1)- feed through cognition, then Δbehaviour
- rescale by the group calibration ratio to get the next year's frequency
The loop iterates until the maximum behavioural frequency change (share of uptake of ASI behaviour) between two consecutive outer iterations falls below OUTER_TOL.
Role: Main model file containing the full economic structure — declarations of variables and parameters, and the model equations and first-order conditions. For a comprehensive overview of the equations, refer to the Appendix of Corbier et al. (2025). Note that the equations have undergone slight modifications since Corbier et al. (2024, 2025, 2026).
Contents:
- Macro definitions: region selection (JPN only for now), SSP scenario, and definitions of sectors, materials, and lifestyles.
- Users can select the region in
MODEL_selection.inc(@#define REGION): Japan (JPN) is available now, with the EU27 and China coming soon. - Users can currently select SSP2, as the LIFE model is only calibrated on SSP2.
- The parameters governing lifestyle changes are the "modifiers". A baseline run holds all modifiers at 0. This repository currently includes the sufficiency and sharing lifestyles; the repair lifestyle is available on demand (darius.corbier@cmcc.it) and will be released by the end of the year.
- Users can select the region in
- Endogenous variables (300+): firms' decisions (output, capital, labour, energy and material inputs); household variables by lifestyle (consumption, saving/investment, energy-using and other durable goods); government (budget, taxes, subsidies); international trade; material stocks and flows; waste flows; CO₂ emissions.
- Predetermined variables: capital stocks, energy-using and other durable goods stocks, material stocks.
- Exogenous variables: technological efficiencies, resource prices, fiscal policies.
- Parameters: substitution elasticities, preference parameters, depreciation rates.
- Model equations: household behaviour (by lifestyle: low-carbon is the highest-income group, cautious is the moderate income group, constrained is the lowest-income group); sectoral production functions (consumer and capital goods, materials, repair and sharing services); market-clearing conditions; physical constraints (material balances, emissions); trade with Armington CES.
Features:
- Uses the Dynare preprocessor with
@#forloops to auto-generate equations. - Perfect-foresight solution, with or without anticipation errors, over 82 periods (2018–2100).
Role: Top-level MATLAB driver: execution and post-processing.
Execution steps:
- Setup — select mode (
'classic'single run, or'calibration'used by the coupling pipeline), model (B2C / C2C), and SSP scenario (SSP2 available). - Generate calibration files —
CIRCEE_calibration.m(structural parameters fromcalibration.csv);CIRCEE_baseyear_values.m(2018 base-year values);CIRCEE_shocks.m(exogenous shock paths fromshocks.csv);CIRCEE_endvalues.m(terminal values for convergence). See section 3.1.2 ofCIRCEE_RunFile.mfor adding exogenous/policy variables. Advice: avoid too many shocks and be cautious with magnitudes to prevent infeasibilities and corner solutions. - Run Dynare —
dynare CIRCEE_PF.mod; compute the steady state fromCIRCEE_steadystatemodel.m; solve the transition path. - Post-processing — rescaling (per-household values scaled to economy-wide totals by population × labour productivity × group population share); unit conversion (energy in EJ, employment in millions); GDP in MER and PPP; CSV generation, including per-scenario level files, per-household variables (
*_percapitain the code), carbon/waste/material footprints, and thematic CIRCEE-LIFE files. - Welfare and distributional analysis — see below (post-processing on saved outputs; no Dynare re-run required).
Role: Analytical steady-state computation. If you modify any equation in CIRCEE_PF.mod, adjust the corresponding steady-state resolution here. Recursive solution from normalised ratios to absolute levels, covering energy prices (from WITCH), lifestyle-weighted parameter aggregation, the numeraire, CES price aggregators, investment, production sectors, household consumption, market clearing, employment, aggregates (GDP, emissions, material flows, waste), the government budget, and trade.
The welfare pipeline runs as post-processing on the saved model outputs, reading the raw per-household simulation stored in the welfare .mat files. It does not require re-running Dynare, so it completes in minutes.
CIRCEE_WelfareAnalysis.m— computes lifetime consumption-equivalent variation (CEV) by lifestyle group, and between-group inequality for energy-service access, consumption, and carbon/waste/material footprints. Inequality is reported as a between-group Gini, a high-to-low per-household ratio, and a between-group Atkinson index (ε robustness). All inequality measures are computed on per-household quantities. With three lifestyle groups, decile-based measures such as the Palma are not well-defined; the high-to-low per-household ratio is reported instead. Also note that, since we do not observe the full income distribution, the Gini and Atkinson results should be interpreted with caution.CIRCEE_WelfarePostProcess.m— the driver for a single run: loads the saved welfare.matfiles, calls the analysis, and exports results. Point it at your own input/output folders via its arguments.CIRCEE_WelfarePostProcess_batch.m— an example driver that loops over lifestyle × scenario × foresight combinations and calls the single-run driver for each. It is provided as a template: edit theUSER PATHSblock at the top and the lifestyle/scenario lists to match your own runs.CIRCEE_WelfareExport.m— exports the welfare and distributional results to CSV.Footprints.m— computes the material, waste, and carbon footprints used in the distributional analysis.
To run the welfare post-processing for a single run:
clear functions % ensure edited functions are reloaded
CIRCEE_WelfarePostProcess('folder', '<path-to-welfare-inputs>', ...
'baseline', 'Baseline', ...
'output', '<path-to-output>')To reproduce a full set of runs, edit the USER PATHS block in CIRCEE_WelfarePostProcess_batch.m and run it; it loops over the lifestyle and scenario lists you specify.
- Columns:
Variable,Scenario,Model,Value. - All structural parameters and most exogenous variables, filtered by SSP scenario and model type. Example parameters: substitution elasticities (σ), distribution shares (α), depreciation rates (δ), fiscal parameters.
- Columns (semicolon-delimited):
Variable;Region;Year;Value. - Time paths (2019–2100) for: energy and material efficiencies (WITCH); share of each lifestyle in total population; resource-price evolution (WITCH); R&D, energy-efficiency, and power-generation investments (WITCH); behavioural modifiers; emission factors.
Two copies exist and serve different purposes:
data/<REGION>/raw/shocks.csv— read directly byCIRCEE_RunFile.min classic mode. Shipped with all behavioural modifiers set to zero, so a fresh clone reproduces the reference run.data/<REGION>/templates/shocks.csv— the skeleton used by the coupled pipeline.build_shocks_csvinscripts/lib/common.shappends the modifier paths for a given lifestyle configuration and writes the result toresults/grid_point_data/shocks.csv.
Saved at the project root, tagged by scenario name:
| File | Contents |
|---|---|
shocks_final_<scenario>.csv |
converged modifier path passed to CIRCEE |
Sharing_final_<scenario>.csv |
final CIRCEE sharing outputs |
Lowering_Expenditures_final_<scenario>.csv |
final CIRCEE sufficiency outputs |
convergence_<scenario>.csv |
max frequency change per outer iteration |
life_dynamics_<scenario>.csv |
full LIFE frequency state per outer iteration × year |
modifiers_sharing_T0.csv |
T0 sharing modifiers, targets, baselines |
modifiers_expenditures_T0.csv |
T0 sufficiency modifiers, targets, baselines |
| File | Contents |
|---|---|
shocks_baseline.csv |
zero-modifier shocks fed to CIRCEE |
Sharing_baseline.csv |
CIRCEE sharing outputs (no modifiers applied) |
Lowering_Expenditures_baseline.csv |
CIRCEE expenditures outputs (no modifiers applied) |
Additional intermediate CIRCEE files land in grid_point_data/ in both modes. Welfare and distributional CSVs (CEV, Gini, high-to-low per-household ratio, Atkinson) are produced by the welfare pipeline.
By production sector: Y_s (output), K_f_s/L_s (capital/labour), M_s/E_s (materials/energy), El_s/Nel_s (electricity/fuels), p_s (wholesale price).
By household lifestyle: C_h (consumption), D_h/D_lowuse_h (durable stock), X_h/SD_h (nondurables/semi-durables), ES_h (energy services), Inv_d_new_h/Inv_d_repair_h (new/repaired durable investment).
Government: Revenues, Expenses, PS (budget); Carbon_budget, EPR_budget (redistribution).
Trade: IMP_s, EXPORT_s, TB (trade balance), Bonds_foreign.
Environment: CO2; MW/IW (municipal/industrial waste); DMC/DMI (material consumption/inputs); M_stock; Gross_additions_stock/Net_additions_stock; Material_balance (mass-balance check).
The current version supports SSP2; further scenarios are planned: SSP1 (sustainable development), SSP2 (middle-of-the-road, default), SSP4 (inequality), SSP5 (rapid growth), NoGrowth (demographic stationarity).
CIRCEE is under active development. Planned extensions include:
- Supply-side circular-economy strategies — the current model focuses on demand-side, consumer-facing strategies (sharing, repair, sufficiency). Future versions will add supply-side strategies such as recycling and green product design (e.g. durability, reparability, and material choices at the design stage).
- Additional lifestyles — the repair lifestyle will be released open-source (currently available on demand).
- More regions — the EU27 and China are planned, alongside the current Japan (JPN) calibration. However, CIRCEE-LIFE will only have Japan, and potentially China. The EU27 is not planned for CIRCEE-LIFE because of lack of surveys.
- More SSP scenarios — SSP1, SSP4, SSP5, and NoGrowth, beyond the current SSP2.
- A Julia implementation of the model.
- Time horizon: 2018–2100. An extra 20 years (82 periods total) are added to avoid end-of-horizon spikes; discard results from 2081 to 2100. The current CIRCEE-LIFE code gives results from 2018 to 2060. Later periods are discarded. If you wish to run CIRCEE-LIFE after 2060, data on lifestyles are available upon request.
- Frequency: annual.
- Solution method: perfect foresight, with or without anticipation errors (see the Dynare manual).
- Expectations handling: support for expectation errors (
learnt_in). - Convergence: automatic steady-state verification.
If you encounter any issues, please contact darius.corbier@cmcc.it. Comments and suggestions for model development are welcome.
References:
- Emmerling, J., Drouet, L., Reis, L. A., Bevione, M., Berger, L., Bosetti, V., … & Havlik, P. (2016). The WITCH 2016 model — documentation and implementation of the shared socioeconomic pathways (No. 42.2016). Nota di Lavoro.
- Pettifor, H., Agnew, M., & Wilson, C. (2023). A framework for measuring and modelling low-carbon lifestyles. Global Environmental Change, 82, 102739.
- Corbier, D., Pettifor, H., Agnew, M., & Nagashima, M. (2025). Shaping sustainable consumption practices: Changing consumers’ habits through lifestyle changes and Extended Producer Responsibility schemes. Resources, Conservation and Recycling, 217, 108214.
- Corbier, D., Pettifor, H., Agnew, M., & Schlegel, N. (2026). Economic incentives and lifestyle drivers: how they shape consumers' engagement in repairing energy-using consumer goods and their environmental impacts in Japan. Global Environmental Change, 96, 103102.
- Corbier, D., Pettifor, H., Agnew, M., & Drouet, L. (2024). CIRCEE, the CIRCular Energy Economy model: Bridging the gap between economic and industrial ecology concepts. Journal of Industrial Ecology, 28(6)
DOI Badge
DOI10.5281/zenodo.23189968Markdown
