A Synthetic-Grid Tax–Benefit Calculator for Comparative Dynamic Microsimulation Using EUROMOD HHoT

This contribution introduces a tax–benefit calculator designed for use in the microWELT microsimulation platform. MicroWELT is a dynamic, comparative microsimulation model based on harmonised survey data and intended to analyse interactions between population ageing, sociodemographic change, and welfare-state regimes. A defining feature of microWELT is its continuous-time, event-driven architecture, in which individuals and households evolve through labour-market transitions, earnings changes, partnership formation and dissolution, fertility, and household re-composition at irregular times. While this design improves realism for life-course dynamics, it imposes demanding requirements for fiscal modelling, as taxes, social insurance contributions, and cash transfers must be updated whenever a relevant event occurs. Embedding full national tax and benefit systems directly into a continuous-time microsimulation model is costly to implement, difficult to validate, and challenging to maintain across countries and reforms. Moreover, modelling explicit fiscal-balancing or budget-closure rules in forward-looking scenarios would require detailed, country-specific assumptions that are often beyond the scope of comparative research. To address these challenges, we develop an approach inspired by the database-imputation strategy proposed by van de Ven et al. (2025), but based on a synthetic tax–benefit database and EUROMOD’s Hypothetical Household Tool (HHoT). Instead of imputing tax–benefit outcomes from survey-based EUROMOD outputs onto simulated agents via nearest-neighbour or hot-deck matching, we construct a synthetic population that spans labour-market states and income sources, finely gridded income levels, partnership status, and detailed child configurations (numbers and ages). EUROMOD is then applied to this synthetic population to generate consistent policy outcomes for personal income taxation, employee and employer social insurance contributions, and benefit entitlements. The resulting EUROMOD outputs are post-processed into multidimensional parameter tables that can be queried efficiently during simulation. These tables are mapped to National Transfer Accounts (NTA) categories, translating detailed programme outputs into harmonised fiscal aggregates suitable for comparative analysis and scenario development focused on broad policy domains rather than programme-specific reforms. Parameters are expressed as tax and contribution rate schedules and benefit schedules conditional on the synthetic-state descriptors. Within microWELT, gross income components and household characteristics are modelled longitudinally; at each event time, the simulator retrieves the relevant parameters from the tables and applies them deterministically to compute net incomes and transfers. Annual totals are accumulated to support alignment with macro targets and life-course accounting. This synthetic-table architecture substantially reduces runtime by avoiding repeated tax–benefit computations during simulation, avoids confidentiality constraints associated with using survey microdata in external policy engines, and mitigates small-sample problems in survey-based EUROMOD inputs (such as limited representation of rare household types or extreme incomes). By working with NTA-mapped aggregates, the approach also facilitates cross-country scenario analysis at a higher level of abstraction. At the same time, it introduces new validation challenges related to coverage of the synthetic grid, discretisation choices, and consistency of NTA mappings to ensure that baseline tax–benefit relationships are reproduced and that the framework can be credibly used in counterfactual scenarios.