Book of Abstracts

Benefit non take-up refers to situations in which individuals or households do not claim social benefits to which they are legally entitled, due to low expected financial gains or costs associated with claiming, including administrative complexity, time and effort, stigma (Moffitt, 1983). While public debate has often focused on benefit fraud, non take-up is considerably more widespread (Ko and Moffitt, 2022). Low take-up undermines the redistributive capacity of welfare states and limits their effectiveness in protecting households against poverty and economic insecurity (Van Oorschot, 1991; Matsaganis et al., 2008). An alternative interpretation, however, views non take-up as a screening mechanism that contains public expenditure by discouraging claims from less needy households (Nichols and Zeckhauser, 1982). In the UK, existing micro-level studies of benefit take-up are relatively dated and rely largely on cross-sectional data (Blundell et al., 1987; Pudney et al., 2006; Zantomio et al., 2010). As a result, there is little evidence on the longer-term consequences of non take-up, partly due to the lack of longitudinal data linking benefit eligibility to observed outcomes. This study addresses this gap by combining longitudinal survey data with tax-benefit microsimulation, complementing recent work on the determinants of non take-up in the UK (Vella and Richiardi, 2024) by focusing instead on its consequences. The main objective of the study is to examine the consequences of non take-up of means-tested benefits for eligible individuals over time, with a focus on labour market outcomes, poverty, physical and mental health, and subjective wellbeing. The analysis combines the UK Household Longitudinal Study (UKHLS) with the UK tax-benefit microsimulation model, UKMOD. UKHLS provides annual panel data on income, employment, household composition, health, and wellbeing. Embedding UKMOD within a longitudinal survey represents a key methodological innovation, allowing benefit eligibility to be reconstructed consistently over time and take-up to be modelled dynamically. Non take-up is identified by comparing observed benefit receipt, based on self-reported information in UKHLS, with simulated eligibility derived from UKMOD. The analysis covers the main social assistance programmes in the UK, namely Universal Credit and its legacy benefits, Pension Credit, and Child Benefit. Prior research shows that eligible individuals who claim benefits differ systematically from those who do not, with non-claimants typically having higher incomes, higher levels of education, and lower dependency loads. In addition, take-up is subject to state dependency, in the sense that individuals who claim benefits once are likely to continue claiming in subsequent periods, and vice versa. To address this selection, propensity score matching is used to construct comparable groups of eligible claimants and eligible non-claimants based on observed characteristics measured prior to take-up. Outcomes are then analysed using a difference-in-differences design, comparing changes over time between the two groups. In addition, individual fixed effects regressions are estimated, exploiting within-person variation in take-up status across waves.

SimPaths is an open-source framework for modelling individual and household life course events, jointly developed at the Centre for Microsimulation and Policy Analysis and the University of Glasgow (Bronka et al., 2025). The framework is designed to project life histories through time, building up a detailed picture of career paths, family (inter)relations, health, and financial circumstances. The modular nature of the SimPaths framework is designed to facilitate analysis of alternative assumptions concerning the tax and benefit system, sensitivity to parameter estimates and alternative approaches for projecting labour/leisure and consumption/savings decisions. SimPaths builds upon standardised assumptions and data sources, which facilitates adaptation to alternative countries. The presentation will focus on recent developments of the SimPaths framework, centred around: • Introduction of new health and well-being variables (MCS, PCS, life satisfaction) for the UK model • Development of a new wealth module for the UK model • Updated release for the model for Italy • First release for the models for Poland, Spain, Germany, Spain • Development of macro modules (all countries) • Development of migration modules (all countries) • Development of inter-household transfer modules (all countries) • Improved model documentation Strategies for internal validation will also be discussed.

Universal Credit (UC) is the main means-tested benefit in the UK welfare system, supporting low-income individuals and families. UC replaced multiple benefits with a single payment, while introducing strict job search requirements and in-work conditionality. Individuals who are not working and are deemed capable of work are usually required, among other things, to actively look for a job, while claimants who are working but earning below a threshold are required to take steps to increase their earnings, including looking for alternative jobs and increasing work hours. Failure to comply can result in benefit sanctions. Research shows that UC conditionality can have detrimental effects on individual well-being and mental health, while evidence of its employment effects is mixed. In this study, we jointly model the take-up behaviour and labour supply decisions through the lens of a structural random utility model. Individuals anticipate that receiving UC negatively affects their well-being, and job search requirements may reduce the utility they derive from income and leisure. As a result, they might choose not to take up UC even if they are eligible and modify their labour supply accordingly. In this paper, we compare baseline simulations with estimated parameters with counterfactual simulations where the effects of conditionality are muted/removed. This allows us to quantify the impact of conditionality on a number of outcomes of interest, including benefit take-up and employment.

We propose a new method for generating replenishment cohorts in dynamic microsimulation models. Standard dynamic microsimulations project the future states of an initial population through the recursive application of one-step-ahead predictions. Over time, sample size declines due to attrition (e.g., mortality), and without the integration of new individuals, the projected population progressively departs from the target population structure. To preserve representativeness, replenishment cohorts must therefore be introduced at each simulation step. Cohort replenishment is challenging because it must simultaneously (i) reflect secular trends in individual characteristics (e.g., declining smoking prevalence) and (ii) preserve the underlying correlation structure among these characteristics (e.g., the relationship between smoking and lung cancer). Existing approaches, most notably the method used in the Future Elderly Model, address these challenges but are computationally intensive and algorithmically complex. We introduce an alternative algorithm that draws eligible donors from historical data while preserving their observed characteristics. Donor sampling weights are adjusted to match period-specific target prevalences using a procedure akin to entropy balancing (Hainmueller, 2012). The proposed method offers three key advantages: (i) efficiency, as it is substantially simpler to implement and less computationally demanding than existing approaches; (ii) accuracy, as it closely tracks specified feature trends given a sufficiently rich donor pool; and (iii) parsimony, as it avoids the need to explicitly specify trends for all variables. Instead, trends in non-targeted characteristics emerge endogenously from the imposed constraints, an especially valuable property in settings with limited longitudinal data. We evaluate the proposed method by comparing its performance with the benchmark approach in reproducing target prevalences, preserving the joint distribution of individual characteristics, and generating plausible trends for features not explicitly constrained features.

Over 15 years since the EU integration of Bulgaria an ambitious reform has been introduced in 2023 regarding the minimum income protection scheme operated within the social policy mix, namely the Monthly Social Assistance (MSA) benefit. The reform addressed European Council’s country specific recommendations (CSRs) to Bulgaria within the framework of the European Semester and had several ambitious goals, among which to improve benefit coverage, adequacy, and targeting. The long years maintained MSA scheme had limited reach due to many strict eligibility criteria so the reform aimed in expanding the inclusion of more vulnerable individuals and families. Besides, the monetary values of the benefit were previously not indexed to inflation as well as outdated means-test thresholds were implemented. From this point of view, the 2023 reform sought to tie the benefit amounts to the national poverty line in order to make them more responsive to the rapidly shifting economic circumstances since the 2022 energy crisis and related inflationary pressures. Moreover, the introduction of annual adjustment aimed in improved reflection of MSA regarding the needs of different groups (e.g., elderly, disabled, single parents). The microsimulation analysis of such effects expected to be achieved by the 2023 MSA reform is performed for the period 2023-2025 using the Bulgarian component of EUROMOD: the tax-benefit microsimulation model of the EU. Most up-to-date dataset is utilized, derived from SILC 2023 survey measuring the incomes for year 2022. Particular indications for the shifts in selected poverty and inequality indicators are evaluated, having in mind that the reform was intended to narrow the poverty gap and improve social inclusion.

What do you want to teach? This hands-on tutorial introduces participants to PolicyEngine, a free, open-source microsimulation platform for analysing tax and benefit policy reforms in the US and UK. Participants will learn to use PolicyEngines web interface (policyengine.org) to: (1) model a tax or benefit reform by adjusting policy parameters, (2) compute hypothetical household impacts showing how the reform affects a hypothetical households taxes, benefits, and net income, (3) run population-level microsimulation analysis to estimate budgetary cost or revenue, distributional effects across income deciles, poverty impacts, and winner/loser breakdowns, and (4) use PolicyEngines AI assistant (a Claude Code plugin) to conduct policy analysis from natural language prompts, including generating charts, policy briefs, and congressional district or constituency-level breakdowns. The session will use live examples relevant to current policy debates in both the US and UK. Why is it useful? PolicyEngine is used by governments (including No. 10 Downing Street), think tanks (Brookings Institution, CRFB, Niskanen Center), and researchers for rapid policy analysis. Unlike proprietary microsimulation models, PolicyEngine is fully open-source and requires no software installation—all analysis runs in the browser or through a Python package. This makes it accessible to researchers, students, and policy practitioners who need to evaluate reform proposals but lack access to established microsimulation infrastructure. The AI-assisted workflow further lowers the barrier, enabling users to conduct distributional analysis and generate publication-quality outputs from plain-language policy descriptions. What is your expertise? Max Ghenis is co-founder and CEO of PolicyEngine. He previously founded the UBI Center, a think tank researching universal basic income policies, and worked as a data scientist at Google. He holds a masters degree in Data, Economics, and Development Policy from MIT and a bachelors degree in operations research from UC Berkeley.

Spatial dynamic microsimulations probabilistically project geographically referenced units with individual characteristics over time. Like any stochastic projection method, their outcomes are inherently uncertain and sensitive to multiple factors. In discrete time dynamic microsimulations, for each simulated time step (often years), each unit passes through different modules addressing different life events (births, deaths, ageing, partnership, employment, …), evaluating if a status transition occurs for them via a Monte Carlo experiment. This inherently introduces uncertainty due to the methods stochastic nature. However, simulations may also be sensitive to other factors, such as the choice of model types and complexity, as well as the parameter estimations, among others. Few articles detail the uncertainty in dynamic microsimulations, and the importance of its components is often overlooked. This is due to the high computational effort required for testing numerous simulation configurations and individual runs are necessary for the analysis. A complete sensitivity analysis testing the sensitivity to each single parameter in every module of the simulation would be unfeasible due to the complex structure of these microsimulations and the resulting computational power required to run them. Moreover, since dynamic microsimulations are typically developed to address specific problems and vary significantly in design and complexity, one-size-fits-all solutions are unattainable. Lastly, there is no commonly agreed-upon standard for reporting uncertainty in dynamic microsimulations. Applying variance-based sensitivity analyses to both direct and indirect effects within the employment module of the MikroSim model for Germany, we show that commonly considered sources of uncertainty, namely coefficient and parameter uncertainty, are less influential than qualitative modelling choices. Dynamic microsimulations being inherently complex and computationally intensive, it is crucial to consider potential factors of uncertainty and their influence on simulation outputs in order to more carefully design simulation setups and better communicate results. We find that simple summary measures do not adequately capture overall model uncertainty and therefore urge modellers to account for these broader sources when designing microsimulations and interpreting their results.

The Department for Work and Pensions (DWP) is responsible for welfare, pensions and child maintenance policy. It administers the State Pension and a range of working age, disability and ill health benefits to around 20 million claimants and customers. This makes it directly responsible for over £300 billion of expenditure every year (equivalent to €340bn or $410bn), as well as providing the regulatory framework for over £2 trillion of private pension assets. People spend most of their lives either paying into a pension or benefitting from it, so the impact of any policy reform unfolds over several decades. The modelling used to provide the evidence base for reform needs to cover that timescale and take in to account a wide variety of life factors including employments, benefit claiming, pensions (accumulation and decumulation), wealth, health, disability, births and deaths. The Department’s current long term dynamic microsimulation model, Pensim3, has underpinned the evidence base for every major state and private pension reform over the past two decades. However, new administrative and survey data sources, alongside improved microsimulation techniques, have created opportunities for a modernised approach. In response, DWP has undertaken an ambitious programme to build a new dynamic microsimulation model, SimPLE, to simulate the entire Great Britain population on an annual time step. The model integrates multiple data sources at the starting point and implements regression-based rules to project long term outcomes. SimPLE is nearing completion, with work focused on finalising priority components ahead of an extensive quality assurance phase. Throughout development, the team has varied in size—from a single modeller to a multidisciplinary group of eight—with staff rotating through promotion, career development and departmental moves. Team members brought mixed levels of prior modelling experience: some with substantial microsimulation expertise, some with general analytical experience, and some entirely new to analytical work. This created challenges in managing the build process and highlighted the importance of: Knowledge management — clear documentation, code comments and development logs. Model modularity — common design principles, consistent coding practices, and a single shared model structure. Training and capability-building — equipping staff with the necessary modelling skills while recognising specialist strengths. Extensive quality assurance — including structured code reviews, testing and validation of outputs. While these practices are valuable in any setting, they are especially important in a government department, where teams are multidisciplinary, turnover can be high, and institutional knowledge must be preserved. This presentation will share our experience of developing SimPLE within this environment, outlining the methods used, the challenges faced, and the steps taken to build long-term modelling capability inside government.

The Material Need Benefit (MNB) constitutes the core component of the Slovak social assistance system. It is a means-tested transfer whose eligibility depends on household composition, income thresholds, and asset tests. In the TATRASK model, which is a static microsimulation model of the Slovak tax and transfer system based on linked administrative data collected by government authorities, the MNB is simulated through the application of legislative rules subject to data availability constraints. As is common in microsimulation models of this type, both the number of simulated beneficiaries and the aggregate fiscal cost are substantially overestimated. This limitation stems primarily from the inability to model benefit take-up accurately due to missing or unobserved information in the data. To address this issue, first an expert-based adjustment approach is implemented, where the number of potential beneficiaries is reduced through a set of rules reflecting observed behavioural patterns in the data. While this method improves model performance and reduces overestimation, it remains limited in its ability to fully capture take-up behaviour. As an alternative, this contribution applies machine learning techniques to model MNB take-up. Specifically, XGBoost models are trained to predict the probability of benefit receipt and to identify likely beneficiaries within the underlying dataset. The results demonstrate that the machine learning approach clearly and substantially outperforms both the baseline TATRASK simulation and the expert rule-based adjustment in terms of accuracy. Moreover, cross-temporal validation confirms the robustness and stability of the ML model, even in the presence of policy changes affecting MNB eligibility.

This paper proposes a novel approach to modelling labour supply by integrating discrete-choice frameworks with search and matching frictions, effectively capturing both supply-side heterogeneity and demand-side constraints. By combining Random Utility-Random Opportunity models with a static search and matching framework, the model offers a realistic representation of involuntary unemployment and wage adjustments, addressing key limitations in the existing literature. Empirical analysis using data from selected EU countries demonstrates the model’s capability to reflect both supply and demand responses. The introduction of an in-work benefit scheme reveals substantial cross-country variation in labour supply reactions and fiscal outcomes, highlighting the critical role of demand-side adjustments in shaping policy effectiveness.