Model Parameterisation

The UK model combines estimated behavioural relationships, demographic and alignment targets, tax-benefit donor data, and scenario assumptions. A simulation's specification depends on both its code revision and the input files supplied to it.

The repository input directory is the canonical file inventory. The parameter-loading code identifies the workbooks and worksheets used by the implementation. The existence of a file alone does not establish that every run uses it: country, module switches and scenario settings also matter.

Tax and benefit inputs

SimPaths imputes tax-benefit outcomes from precomputed UKMOD donor files in input/EUROMODoutput/. The policy schedule maps simulation years to policy systems. Tax-Benefit Donors (UK) explains how to obtain and prepare the files; Modifying Tax-Benefit Settings explains scenario configuration.

Training donor data are available in the repository for testing. They are not a substitute for authorised research data.

Parameters and targets

File family	Role
reg_*.xlsx	Regression coefficients and, where provided, covariance matrices. Some workbooks contain multiple processes; reg_RMSE.xlsx contains residual error scales rather than a separate behavioural equation.
align_*.xlsx and alignment_targets_*.xlsx	Targets used to align selected simulated outcomes.
projections_*.xlsx	Demographic projections, including fertility and mortality inputs.
scenario_*.xlsx	Time paths and assumptions for the selected scenario.
EUROMODpolicySchedule.xlsx	Maps simulation years to tax-benefit systems.
social_care_parameters.xlsx	Social-care policy parameters.
economic_time_series.xlsx	Price and wage indices, returns, debt costs and care wages.
alignment_adjustment_series.xlsx	Time-varying adjustments for alignment processes.

The Repository Guide's estimation reference maps UK Stata scripts to the outcomes they estimate. This is distinct from the workbook inventory: a script can update several workbooks, and a workbook can supply several processes.

For example, reg_health.xlsx supplies self-rated health and disability; reg_health_wellbeing.xlsx supplies SF-12 physical and mental health scores and life satisfaction; and reg_financial_distress.xlsx supplies financial distress. These should not be treated as interchangeable health specifications.

Interpreting estimates and references

The SimPaths reference paper provides the model-level description. The Labour supply and Mental health pages explain the corresponding processes.

Richiardi and He's 2021 paper, “No one left behind: The labour supply behaviour of the entire Italian population”, concerns the Italian application. It should not be read as an inventory of the current UK coefficient estimates. For UK work, consult the labour-supply specification and the parameter files for the revision being used.

Kopasker et al. (2024), Evaluating the influence of taxation and social security policies on psychological distress, describes a UK application. Later development versions can change its implementation or extend the outcomes; the paper and current code are not automatically identical.

Reproducible use

Record the code revision, parameter-file versions, initial population, donor data, policy schedule and configuration with each analysis. After re-estimation, check coefficient names, outcome coding, worksheet names and covariance matrices against the Java loader, then perform the relevant validation. Do not mix parameter files from different revisions without checking compatibility.

See Uncertainty Analysis for the distinction between coefficient resampling, simulation randomness and sensitivity to assumptions.