Blocks

Blocks are reusable building blocks that emit equation ASTs and model variables. A model is assembled by selecting and configuring blocks, then organizing them into sections (production, households, markets, etc.). Each block has a clear responsibility and declares which sets and parameters it requires.

Forms

Many blocks accept a form parameter to choose the functional form (for example Cobb-Douglas or CES). Use a global form for all items or provide a per-entity mapping where supported.

Typical forms include:

• Cobb-Douglas for smooth substitution.
• CES for flexible elasticities.
• Leontief for fixed proportions.

Typical block categories

Most models mix a small set of block categories:

• Production blocks that map inputs to outputs.
• Market clearing blocks for goods and factors.
• Income and demand blocks for institutions.
• Closure and numeraire blocks.
• Auxiliary-quantity blocks for calibrated quantities that must enter model equations but are not monetary SAM accounts.

Organize blocks into sections so the model structure is readable and easy to diagnose.

Parameterization

Blocks take parameters from calibrated data. Keep parameter names consistent across blocks so calibration outputs can be reused. Prefer explicit mappings over implicit defaults when building large models.

Auxiliary quantities

quantity_link, quantity_transformation, quantity_balance, and quantity_capacity add a small, general algebra for supplementary quantities that must participate in equilibrium conditions. They are intentionally domain- and unit-agnostic: mappings define relationships, while coefficients and capacities are supplied by the model's calibration data.

Use a link to connect a quantity to an existing model variable, a transformation for coefficient-based conversion, a signed balance for an identity, and a capacity for an upper limit. Inputs must be defined by an earlier block. Use these blocks only when the quantity changes the equilibrium system; use JCGEOutput satellite reporting when the quantity is reported from solved model drivers without adding an equilibrium condition.

Extending blocks

Custom blocks can be created by following the same interface as the built-ins: declare required sets, map parameters, and emit equation ASTs. This makes new components fully compatible with the runtime and output tooling.

Guidelines

• Blocks should be reusable and model-agnostic.
• Avoid hard-coded sector/factor counts; always work with sets.
• All equations are registered with block and tag metadata to enable rendering and reporting.
• closure_condition(block, tag, indices...) constructs the stable key for an equation when a model needs to mark it as an accounting check.

See Closures & Checks for the role declaration and runtime behaviour.

See the JCGEBlocks documentation at https://Blocks.JCGE.org for the full block catalog and parameters.