Model configuration YAML

model_configuration.yaml describes one model version and build. It lives in the model-build root, for example 6W/200/model_configuration.yaml.

This complete configuration identifies the model, its years and source data mappings, and explicitly selects the common growth rate:

schema_version: 1

model:
  version: "6W"
  build: "200"
  sector_count: 6

years:
  base: 2023
  last_projection: 2150

mapping_columns:
  gtap:
    country: "W"
    commodity: standard_6

calibration:
  effective_labor_productivity_growth_rate_in_steady_state: 1.4

Quote model.build. The value is a string even when it contains only digits. The version and build must agree with the directory containing the file.

Growth units and compatibility

The following partial mapping sets 1.4 per cent annual log frontier growth:

calibration:
  effective_labor_productivity_growth_rate_in_steady_state: 1.4

The default is also 1.4. Advancement CSV cells are percentage-point deviations from this rate, with exact zeros required for the final 50 configured projection years. The Python configuration property remains a decimal, so this YAML produces 0.014 internally.

Build 200 introduces this interpretation. Build 199 and earlier used decimal-valued configured growth and absolute percentage rates in the CSV. schema_version remains 1; the build 200 change does not introduce another schema version. Existing decimal-valued configurations and absolute-rate CSVs must be migrated together; old releases require their original input contract. There is no magnitude-based or build-number-based format detection. See calibration settings and migration.

Syntax and recipes

The packaged model_configuration-v1.schema.json file is the authoritative grammar. G-Cubed also performs checks that depend on the selected model, its SYM declarations and its files.

Table of contents