Research project B2/191/P3/BABEL (Research action B2)
PROJECT DESCRIPTION
Context and research objectives
A basic income (BI) is a radical departure from traditional public welfare provision as it severs the link between contribution and benefit on the one hand and between need and benefit on the other hand. Although a BI is often presented as a simple idea, discussing concrete policy proposals quickly ends in a Tower of Babel-like argument. In many of these debates, it appears that proponents, and respectively opponents, often have different kinds of BI schemes in mind when advocating for or criticizing the idea. It is, however, quite obvious that the actual outcomes of a BI will be highly dependent on the concrete policy design in terms of entitlement, eligibility criteria, benefit levels, financing, and implementation trajectories. These aspects will substantially influence the extent to which a BI may discourage paid work or any work, increases or reduces gender inequalities in work and care, and end the myriad of problems associated with means-tested social benefits. Unfortunately, for many of these outcomes we lack empirical scrutiny. This is the main focus of the Basic income in BELgium (BABEL) project.
The BABEL project aims to narrow the gap between claim and reality with regard to BI outcomes. The project has several objectives. First, it aims to investigate the actual labour supply effects of BI. To this end, a quasi-experimental approach is applied in which we use administrative records
to estimate the labour supply effects of BI based on what gets closest to a real world basic income situation: the Belgian Win-for-Life lottery. Second, to gauge the potential outcomes of BI and its different policy versions a comprehensive microsimulation exercise is conducted to estimate the first-order income distributive and budgetary effects of a set of BI proposals and their effect on work incentives in Belgium. Third, by carrying out factorial vignette experiments, the project aims to gain deeper insight into public support for the implementation of a variety of basic income schemes in Belgium, and whether and to what extent public support of basic income schemes depends on the outcomes, financing and implementation details. Fourth, BABEL will pay due attention to the on-the-ground implementation and the technical and administrative feasibility of a selection of basic income policy proposals, and it will gauge to what extent political parties and social partners as gatekeepers in the Belgian welfare state are willing to support the implementation of these schemes. Finally, the project will present a blueprint of pathways for basic income policy proposals that (1) are likely to garner sufficient support by the general public and by social partners; (2) lead to better outcomes in terms of social protection and work incentives; and (3) can inspire feasible welfare reform in Belgium.
Methodology
Work Package (WP) 2. The consequences of implementing basic income: a microsimulation exercise
Using the most accurate microsimulation modelling available, we will assess first-order redistributive and fiscal effects of various BI proposals (selected in WP1) as compared to the current baseline. We will draw on EUROMOD, an advanced microsimulation model that captures the policy details of Belgian cash benefit entitlements and direct taxes and social insurance contributions. The model runs on the EU-SILC, a representative sample of households. As a static model, EUROMOD is particularly useful for the analysis of “day after” impacts of a policy reform on levels of poverty and inequality, as well as on aggregate expenditures and fiscal requirements. EUROMOD is also highly useful to gauge the impact of tax/benefit reforms on work incentives since it allows to calculate participation tax rates (PTR) and marginal effective tax rates (METR) for each simulated policy scenario compared to the baseline.
WP 3. Estimating labour supply effects in a real-life setting: the W4L lottery
We will exploit data from the Win for Life lottery where winners receive a lifelong, unconditional monthly cash payment on an individual basis. Using this this natural experiment, we can assess the causal employment effects of receiving a BI. The current 650 winners will be matched to administrative panel data on the socio-economic trajectories of these individuals, as well as with basic information on the members of their household. We also draw the same data for a control group. We then apply a so-called differences-in-differences estimator (DID) to estimate the difference in employment trajectories between control and treatment group. We collect data for treatment and control groups at one year before (T0) and six months (T1) as well as two years after (T2) winning W4L to allow some time for adjusting behaviour. The variation in the amounts can also be exploited to test whether higher or lower amounts affect labour supply in different ways. The analyses are controlled for the year of winning and for basic individual and household characteristics such as sex, age, number of children, marital status and level of education. We will also be able to check for household effects (labour market behaviour of secondary earners). To account for participating in lotteries being socially stratified and vulnerable for selection effects, a second control group will be carefully selected using Propensity Score (PS) with one-on-one matching. DID and PS are complementary methods to account for potential selection bias.
WP 4. Understanding popular support for BI: the factorial vignette study
To map the multidimensionality for all selected BI schemes (cf. WP1) in terms of different characteristics, outcomes and modalities, we will develop and implement an online factorial vignette survey. Respondents are confronted with descriptions (vignettes) of hypothetical but realistic BI scenarios in which both the policy design and the policy outcomes are randomly varied across vignettes. Each respondent will be required to read several of these vignettes and to rate the respective BI scenarios on a scale ranging from complete opposition to full approval. This way, we can examine whether experimental manipulations of the features and outcomes of a particular BI policy affects the level of support for it; how different BI dimensions interact with people’s background characteristics, and whether some sociodemographic or ideological groups react more strongly to certain dimensions of a BI policy. Respondents will be recruited through an online sample of the Belgian population. Additionally a specific sub-sample of trade union members will be targeted. The number of vignettes per respondent and its associated power will be determined on the basis of the realised sample size.
WP 5. The nuts and bolts of welfare reform: institutional analysis
In WP5, we identify the political constraints and opportunities to actually implement (a variety of) BI scenarios in the Belgian welfare state. In doing so, we draw on qualitative approaches, including semi-structured interviews and focus groups with key players such as civil servants knowing the details of particular benefits, trade unions and employer organisations as gatekeepers of the social security system, and representatives of political parties. We also use text analyses of secondary sources such as internal reports, working papers, pamphlets and memoranda of political parties and the social partners. Finally, we compile a new database (BABELIS) with detailed information on the characteristics of benefits in the Belgian welfare state and to what extent these benefits deviate from BI principles. This is done to estimate the political and technical costs of transition towards a BI in Belgium, bearing in mind the multilevel structure of the Belgian welfare state with regional, local and federal competencies in these domains.
Potential impact of the research
In an integrated way and with a strong valorisation by scientist as well as key players in the Belgian welfare state, the BABEL project will explicitly inform policy-makers about the labour supply effects of different BI policy proposals, their effect of lifelong learning and other decision in modern labour markets. It also informs about the effects of a BI on poverty and inequality levels, on aggregate expenditures and fiscal requirements as well as on what influences public support for a specific BI policy proposal over another. Finally, it will provide pathways for which BI schemes are supported in the Belgian context and how they could be implemented in the existing welfare arrangements. More implicitly, the project will deliver innovative and inspiring research results that can serve other research and experiments on BI in other contexts
