Assessing the effectiveness of Minimum Income (MI) schemes in poverty alleviation is challenging. Studies based on survey microdata are usually subject to bias because households tend to underreport benefit receipts. Studies based on microsimulation models tend to overestimate these benefits mainly due to high non-take-up rates. In this paper, we tackle these challenges by calibrating the simulation of MI schemes in the microsimulation model EUROMOD to match official expenditure results for each European Union (EU) Member State's scheme. We use this calibrated model to evaluate the poverty-alleviating effects of existing MI schemes and explore the impacts of possible reforms toward eradicating extreme poverty. Our results show that MI support varies widely across EU countries, with most failing to reach half of households in extreme poverty or provide adequate support, particularly for larger families. Reforms aimed at improving coverage and adequacy could eradicate extreme poverty at a relatively low budgetary cost.

1 Introduction

Minimum Income (MI) schemes play a crucial role in alleviating poverty and social exclusion in the European Union (EU), constituting a last-resort safety net for households with insufficient resources. While all EU countries provide some form of MI support (see Natili 2020 for a typology of MI schemes in the EU), the effectiveness of these schemes in reaching those in need is highly heterogeneous across countries (Coady et al. 2021; Figari, Matsaganis, and Sutherland 2013; Frazer and Marlier 2016; Immervoll et al. 2022; Nelson 2013). This heterogeneity typically reflects the interaction between two main dimensions: (i) coverage, that is, the extent by which support reaches individuals in need, and (ii) adequacy, that is, the generosity of the support provided. Recently, the European Commission has proposed a Council Recommendation to improve, among other features, these two dimensions of MI support (European Commission 2022a).

Comparing the effectiveness of MI schemes across countries poses several methodological and data-related challenges. Some studies perform a somewhat descriptive analysis based on institutional country-level data reporting benefit levels for hypothetical family types (Marchal, Marx, and Mechelen 2014; Nelson 2010; Wang and Van Vliet 2016). While this approach provides simple and comparable results on the heterogeneity of MI provisions across countries, it does not allow for an assessment of their distributional effects, given the non-granularity of the data. Other studies, while also adopting a descriptive approach, take a step further, using self-reported survey microdata (Ayala and Bárcena-Martín 2020; Coady et al. 2021; Immervoll et al. 2022) or administrative microdata (Bargain, Immervoll, and Viitamäki 2012). Administrative microdata, although more comprehensive and precise in nature, are often not easily accessible, and lack cross-country comparability. In contrast, survey data are more easily available and typically subject to harmonization procedures. Yet, they usually suffer from bias as households with low incomes tend to under-report benefit receipts (Bargain, Immervoll, and Viitamäki 2012; Figari et al. 2012; Lynn et al. 2004), which leads to an underestimation of the support provided by MI schemes.

An alternative approach is to use microsimulation modeling (Figari, Matsaganis, and Sutherland 2013), which simulates the levels of MI benefits for each entitled unit according to the policy rules in place in each country. This method proves to be useful in addressing the problem of under-reporting of benefit receipt present in survey data (Figari et al. 2012), as it allows for a more accurate identification of the population eligible to receive MI support. Furthermore, it provides a finer assessment of the different existing MI schemes, clearly disentangling them from other similar cash benefits that tend to be aggregated in surveys.¹ However, microsimulation modeling also faces limitations, mostly related to the lack of detailed data necessary to accurately simulate some eligibility conditions, as well as the large extent of non-take-up of these benefits (Ko and Moffitt 2024). This leads to an overestimation of cash benefits (Figari et al. 2012; Figari, Matsaganis, and Sutherland 2013; Jara and Tumino 2013; Sutherland and Figari 2013), particularly MI schemes, reflecting their “intended” effects rather than their actual impact.

In this paper, we explore a new approach to the comparison of MI schemes across countries, which builds on the advantages of microsimulation models while trying to overcome the limitations of this approach, using the methodology developed by Hernández, Picos, and Riscado (2022). We use EUROMOD, a static tax-benefit microsimulation model for all EU countries, and apply a calibration procedure that corrects for the overestimation of MI schemes by matching the simulated levels of expenditure to official recorded expenditure data from administrative sources. Although ad-hoc corrections to the simulation of MI schemes have been explored from a country-specific perspective (e.g., Tasseva 2016) or for a selection of EU countries (e.g., Matsaganis, Paulus, and Sutherland 2008), this is, to the best of our knowledge, the first time that such assessment has been conducted for all EU countries using a harmonized approach. Having obtained a new “closer to reality” baseline, we then provide new estimates of the effectiveness of existing MI schemes, focusing on their levels of coverage and adequacy and their poverty-alleviation effects. Furthermore, we explore possible reforms to existing schemes, simulating a number of stepwise scenarios with the aim of eradicating extreme poverty in the EU and estimating the corresponding “morning-after” cost.

Our paper contributes to the literature on the evaluation of MI schemes in four main ways. First, it provides a new methodological framework, which departs from previous microsimulation studies focused on the “intended” effects of MI schemes (e.g., Figari, Matsaganis, and Sutherland 2013) by introducing a calibration procedure to correct for the overestimation of benefit recipiency. The new methodology not only produces a “closer to reality” simulation of existing MI support in each EU country, but also constitutes a harmonized framework that can be used for comparisons in a consistent way. This is key for public policy purposes, as it increases the accuracy and comparability of MI support and poverty measurement across EU Member States. Second, the paper introduces an innovative framework for identifying existing gaps and reflecting on possible ways to reform current MI schemes, sharing useful insights about the potential and challenges of different options, which is also a crucial input from a policy perspective. Third, the paper presents updated results for all EU countries (the last cross-country estimates were produced more than 10 years ago for 14 EU countries, in Figari, Matsaganis, and Sutherland 2013). Having up-to-date estimates is crucial for EU policymakers, to identify existing gaps and understand in which dimensions there is still room for improvement. Finally, the paper delivers estimates of coverage and adequacy using both stricter and broader definitions of MI protection. By doing so, we shed new light on the sensitivity of results to the definition of MI schemes.

The rest of the paper is organized as follows. In Section 2, we describe the concept and main characteristics of MI schemes in the EU. In Section 3, we present the way these schemes are simulated in EUROMOD and discuss some limitations. In Section 4, we introduce our calibration procedure and provide some indicators to validate the new calibrated baseline simulation. In Section 5, we assess the effectiveness of existing EU MI schemes in poverty alleviation based on the new calibrated baseline simulation. In Section 6, we describe the hypothetical reform scenarios and present the impact that they would have on the anti-poverty effectiveness and cost of MI schemes in the EU. Finally, Section 7 discusses the main results, limitations of our approach, and avenues for future research.

2 Concept and Main Characteristics of MI Schemes in the European Union

As described by European Commission (2022a), the provision of MI protection encompasses a large variety of instruments, namely benefits in cash or in-kind, social services, and tailored assistance. Finding an appropriate and comparable definition for a MI scheme is a challenging task. In its broader definition, this would be a cash benefit that helps guaranteeing a MI floor, implying that several schemes would fit such objective. As pointed out by Casas (2005), “there is a wide degree of divergence around the understandings and definitions of this MI floor in the EU.” The author illustrates this by presenting how the different country-specific schemes refer in some cases to the concept of “minimum” (e.g., “minimum living income” in Spain or “guaranteed minimum income” in Greece), whereas in others, these are labeled without any explicit allusion to this concept (e.g., “subsistence benefit” in Estonia or “benefit in material need” in Slovakia). These heterogeneous understandings are also reflected in the literature, with different studies considering wider or stricter classifications of MI schemes. For example, in Spain, Ayala et al. (2021) broadly review all means-tested non-contributory benefits in their analysis of the effectiveness of MI protection, including, for example, unemployment assistance. For the same country, however, Hernández, Picos, and Riscado (2022) consider unemployment assistance an intermediary step before MI support.

While harmonization remains a challenge, and acknowledging that a conceptual discussion goes beyond the scope of this paper, our definition of MI schemes relies on similar grounds as the one in Natili (2020). This broadly means considering MI schemes as anti-poverty cash benefits, with means-tested and non-contributory characteristics, which usually work as a top-up depending on family size and composition. This implies that MI schemes complement all other incomes, including other benefits, up to the MI threshold. These attributes lead us to consider only the “core” MI schemes of each EU country—“core” in the sense of meeting most of these characteristics, and not “core” quantitatively-wise (e.g., expenditure vis-à-vis other benefits).² Furthermore, we focus on schemes targeted at the working-age population, excluding schemes directed only at the elderly and schemes covering special circumstances, such as disability. We typically exclude unemployment assistance schemes as well.³ Following these criteria, we identify the schemes presented in Table 1 as the MI schemes to be considered in the analysis.⁴ We focus on the schemes in place in 2019, a choice we discuss in the next section.⁵ When cross-checking this list with the schemes included in the European Commission's 2019 Mutual Information System on Social Protection (MISSOC) for the working-age population, we find that they largely coincide.⁶

TABLE 1. List of MI schemes considered in the analysis.

Country	MI scheme	Residence requisite	Age requisite	Owner-occupied housing condition	Earnings disregard
		Key features
AT	Guaranteed minimum resources (Mindestsicherung)	Yes	No	Yes	Yes
BE	Integration income (Revenu d'intégration/leefloon)	Yes	Yes	No	Yes
BG	Social assistance allowances (Mесечни социални помощи)	Yes	No	Yes	No
CY	Guaranteed Minimum Income (Eλάχιστo Eγγυημένo Eισόδημα)	No	Yes	No	Yes
CZ	Allowance for Living (Příspěvek na živobytí)	Yes	No	No	Yes
DE^a	Subsistence benefit (Hilfe zum Lebensunterhalt)	Yes	No	Yes	Yes
DE^a	Unemployment assistance for jobseekers (Grundsicherung für Arbeitsuchende)	Yes	Yes	Yes	Yes
DK	Social assistance (Kontanthjælp)	Yes	Yes	Yes	Yes
EE	Subsistence benefit (Toimetulekutoetus)	No	No	Yes	Yes
EL	Social Solidarity Income (KΟINΩNIKΟ EI∑ΟΔHMΑ ΑΛΛHΛEΓΓYH∑)	No	Yes	Yes	Yes
ES^a,^b	Regional Minimum Income Schemes (Rentas Mínimas de Inserción)	Yes	Yes	No	No
ES^a,^b	Minimum Living Income (Ingreso Minimo Vital)	Yes	Yes	No	No
FI	Basic Social Assistance (Perustoimeentulotuki)	No	No	No	Yes
FR^a	Active solidarity income (Revenu de Solidarité Active, RSA)	Yes	Yes	Yes	Yes
FR^a	Employment bonus (Prime d'activité)	Yes	Yes	Yes	Yes
HR	Guaranteed minimum benefit (Zajamčena minimalna naknada)	No	No	Yes	Yes
HU	Benefit for persons in active age (Aktív Korúak Ellátása)	No	Yes	No	Yes
IE^a	Supplementary Welfare Allowance	No	No	No	No
IE^a	Jobseeker's Allowance	No	Yes	No	No
IT	Guaranteed Minimum Income or Pension (Reddito di Cittadinanza)	Yes	No	No	No
LT	Cash social assistance (Piniginė Socialinė Parama)	No	No	Yes	Yes
LU	Social inclusion income (Revenu d'inclusion Sociale, Revis)	Yes	Yes	Yes	Yes
LV	Guaranteed minimum income benefit (Pabalsts garantētā minimālā ienākuma līmeņa nodrošināšanai)	No	No	No	No
MT^a	Social assistance (Ghajnuna Socjali)	No	Yes	No	No
MT^a	Unemployment Assistance (Gℏajnuna gℏal-Diżimpjieg)	No	Yes	No	No
NL	Participation Act (Participatiewet)	No	Yes	Yes	No
PL	Periodic Allowance (Zasiłek okresowy)	No	Yes	No	No
PT	Social integration minimum income (Rendimento social de inserção)	Yes	Yes	Yes	Yes
RO	Social Aid (Ajutor social)	No	Yes	No	No
SE	Social assistance (Ekonomiskt bistånd)	No	No	Yes	Yes
SI	Financial Social Assistance (Denarna Socialna Pomoč)	No	No	No	No
SK	Material Need Benefit (Dávka v hmotnej núdzi)	No	Yes	No	Yes

Source: Our own elaboration based on EUROMOD Country Reports, MISSOC database, and European Commission (2022a, 2022b).
^a When examining countries with two MI schemes, we have attempted, wherever possible, to distinguish the key features of each scheme. In cases where detailed information on both schemes was unavailable, we assumed that the conditions of the primary scheme also applied to the secondary one.
^b In Spain, regional MI schemes are highly diverse across regions. For example, some regions apply earnings disregards, while others do not. For a more comprehensive analysis, see Hernández, Picos, and Riscado (2022).

We acknowledge that our focus could eventually result in analyzing some schemes that are either residual, due to the delivery of previous major insurance-related benefits, or more ambitious because the welfare regimes of certain countries are not predominantly structured around insurance-related programs (see Immervoll et al. 2022 pages 8 and 9 for a detailed discussion). Moreover, by focusing on the “core” MI schemes, we are de facto analyzing only anti-poverty instruments. As such, we generally leave aside complementary sources of support not explicitly targeted at poverty reduction, to which the claimant has to apply separately, such as housing benefits and family allowances, among others. This may imply excluding some benefits that, in practice, have a non-negligible impact on poverty alleviation. For example, although child-related benefits primarily aim at compensating childrearing costs, they can be more effective at reducing child poverty than strict anti-poverty schemes (Bárcena-Martín, Blanco-Arana, and Pérez-Moreno 2018; Leventi, Sutherland, and Tasseva 2019). Consequently, our analysis should not be interpreted as a comprehensive study of the effectiveness of all forms of MI protection in poverty alleviation, but rather as a “scheme-based” analysis. However, we believe our approach has at least two advantages with respect to a more encompassing conceptualization of MI protection. First, it allows for a clear identification of the policy instrument under assessment, which favors more concrete scheme-oriented recommendations for potential reforms. Second, it focuses on evaluating the main goal of these policies, that is, poverty alleviation. If the schemes under assessment do not address to a reasonable extent this goal, even if other schemes do indirectly, policymakers may consider redesigning the benefits system to ensure that MI benefits can be clearly identified as the key anti-poverty tool.

Besides the above-discussed conceptual definitions, MI schemes are characterized by a set of rules. These rules determine the eligibility to access the benefit, the extent to which the benefit can be combined with employment income, whether the benefit is taxable, among others. They can be broadly classified into two categories: non-income related rules and income related rules. Non-income related rules typically pertain to socio-demographic characteristics, asset ownership, and job-search related requirements. Income related rules usually refer to disregards in employment income when computing the amount of the MI benefit, the existence of complementary benefits, and the rules of taxation applied to the MI benefit. In Table 1 we identify whether four of these rules are present in each country, which are the ones that show higher heterogeneity across countries. In Table A4 of Appendix 1 we provide a more comprehensive and detailed list of the most relevant rules.

In most countries, several non-income related rules apply, with countries varying significantly in the degree of restrictions that they impose. For example, while in Italy a minimum of 10 years of residence is needed to qualify for the MI benefit, in Latvia no minimum period is required. Similarly, while in Cyprus only adults aged 28 years or older can qualify, in Austria there are no age requirements. While restrictions on time of residence and age are applied in only a part of EU countries, the vast majority has some type of constraint on asset ownership and job-search efforts. Income-related rules are also present in most countries, with the extent of requirements varying across countries. Typically, a share of employment income is disregarded when computing the amount of the MI benefit, although not in all countries, and one or more complementary allowances may be provided, including housing allowances. In most countries, MI benefits are not subject to income tax.

The spending with MI schemes varies considerably across EU countries (Figure 1). The cost of MI schemes as a percentage of GDP ranges from 0.02% in Latvia to 0.78% in France. The six countries depicting the highest expenditure are France, Cyprus, the Netherlands, Slovenia, Germany, and Denmark, where MI support represents at least 0.5% of the GDP of each country. On the contrary, in Romania, Czechia, Hungary, Bulgaria, Poland, and Latvia, the expenditure in MI schemes accounts for only up to 0.06% of the GDP. The differences in spending are likely to reflect several factors, including the strictness of the eligibility criteria, take-up rates, and the rules applied when calculating the amount of MI benefit for each beneficiary.

Details are in the caption following the image — **FIGURE 1**
Open in figure viewer PowerPoint

Spending with MI schemes, as a percentage of GDP. In countries where two MI schemes are considered, the graph reflects the joint impact of both schemes.
*Source*: Data on MI official recorded expenditure obtained from EUROMOD Country Reports. GDP data retrieved from Eurostat
.

3 The Simulation of MI Schemes in EUROMOD

The analysis in this paper uses EUROMOD, the open-source static non-behavioral tax-benefit microsimulation model for all EU countries.⁷ The tax-benefit systems under assessment are those in place as of June 2019.⁸ EUROMOD uses microdata from the EU Statistics on Income and Living Conditions (EU-SILC). For this paper, we use the 2019 cross-sectional wave, comprising 2018 incomes. Uprating factors are used to bring the reported income values from the income reference period up to the policy year.

The use of microsimulation models for the analysis of MI schemes is relatively ample in the literature. While most studies use microsimulation modeling for estimating the extent of non-take-up (Bargain, Immervoll, and Viitamäki 2012; Fuchs et al. 2020; Matsaganis, Levy, and Flevotomou 2010), which can be substantial in several EU countries (Eurofound 2015), few studies use these estimates for correcting the simulation of these schemes toward a more “realistic” evaluation of their effects (one exception is the study of Matsaganis, Paulus, and Sutherland 2008 for 4 EU countries and the UK). The latter has to do with the challenging task of an appropriate identification of the real beneficiaries. As such, the common practice is to assume full take-up or, more comprehensively expressed, no simulation errors, as in the EU cross-country comparative studies by Figari, Matsaganis, and Sutherland (2013) or in Leventi, Sutherland, and Tasseva (2019).⁹ In this paper, however, we try to reconcile the simulations of the different EU MI schemes with the aggregate numbers collected and reported by each national administration. For this purpose, the first key question we investigate is how accurately the conditions to access these schemes are simulated in EUROMOD, which is crucial to determine the correct identification of potential beneficiaries. In line with the previous section, these conditions can be split into two broad categories: income-related conditions and non-income-related conditions.

Regarding the first category, income tests are generally simulated in EUROMOD with a high degree of accuracy given the availability of detailed income variables in EU-SILC. This means that we account for the different income sources included in the means-test of each scheme, while simulating any differential treatment to those incomes if explicitly designed by the law. However, two caveats apply. First, due to a lack of information about monthly incomes, the income test assumes that the annual income of the assessment unit is equally distributed throughout the 12 months. Thus, possible short entitlements to MI support might not be simulated, eventually leading to undersimulation. Second, to the extent that the detailed income variables in EU-SILC are subject to measurement errors (Trindade and Goedemé 2020), the subsequent simulations would be influenced accordingly. For instance, if an income source is misreported in the survey, this might result in simulating as entitled some units that would be otherwise not eligible. The opposite, although more unusual, could also happen and result in undersimulation. It is difficult to foresee whether the aforementioned issues lead to under or oversimulation, especially if they are simultaneously present in the data.

Regarding the second category, rules related to certain socio-demographic characteristics, for example, age-related requirements and family links between cohabitants, are simulated accurately since they are usually well-captured in EU-SILC. However, other non-income conditions can only be partially simulated. Typical examples include asset-related conditions¹⁰, time of residence in the country, registration at employment offices, or refusal to take up jobs. As a result, the number of beneficiaries in EUROMOD (and hence expenditure) is usually overestimated, that is, the model identifies as eligible some units that would not be entitled to the benefit if non-income conditions were simulated. Additionally, the model overestimates the number of beneficiaries because in reality not all entitled units do take up the benefits due to both demand-side factors (e.g., stigma, lack of information) and supply-side factors (e.g., discretionary rules applied by the administration, budgetary limitations).¹¹ This overestimation has crucial implications when assessing the poverty-reducing effects of MI schemes, as not only the poverty-reducing impact of a country-specific scheme might be overestimated but the ordering of countries in a cross-country comparison might also significantly change depending on the extent of non-take-up (and/or the accuracy of the simulations) across countries.

To tackle the overestimation explained above, some country models in EUROMOD apply a benefit take-up adjustment to remove the beneficiary status to some simulated entitled families. However, this is done only in some countries (Belgium, Spain, Estonia, Finland, France, Croatia, Ireland, Poland, Romania, Slovenia, and Slovakia) and following different methodological approaches, varying from the application of available non-take-up estimates to the simulation of ad-hoc adjustments (e.g., disregarding beneficiaries entitled to small benefits).

4 A Methodology to Calibrate the Simulation of MI Schemes

The shortcomings of simulating MI schemes pose important problems not only for an accurate assessment of the effectiveness of existing MI schemes but also for a meaningful evaluation of possible reforms to these schemes. To do this, which is the focus of the next two sections, we start by applying a calibration procedure that corrects the existing simulations in EUROMOD, by matching the simulated outcomes of MI schemes with official statistics. The calibration is performed in a harmonized way for all EU countries to allow for a meaningful cross-country analysis. We therefore start by removing the ad hoc adjustments that are applied in EUROMOD for a few countries, as mentioned above, to obtain a “clean” baseline, representing the EUROMOD simulations without any adjustment to account for overestimation. We call this the “Full entitlement baseline.” We then apply our calibration procedure to this baseline to match the recorded official statistics and obtain a closer-to-reality “Calibrated baseline.”

4.1 The Choice of External Data

To calibrate the model, three main types of external data can be considered: take-up ratios, number of beneficiaries, and total expenditure. The first option would maintain relative consistency in EUROMOD but in practice, it is not feasible, as take-up rates are not available in a comparable and harmonized way across EU countries. The other two options have the advantage of being consistent with reality in absolute terms, accounting simultaneously for non-take-up and other simulation errors (e.g., the non-simulation of non-income conditions). Using the number of beneficiaries is less reliable than using expenditure, since there may be important conceptual and measurement differences across countries. For example, it is often not clear what the measurement unit reported in official statistics is (e.g., family heads, all individuals in the family or household), nor how they are counted (e.g., in a specific point in time or as an annual average). In contrast, total annual expenditure is a less ambiguous variable, typically corresponding to total spending on the scheme in the year. We therefore use it for the calibration, relying on the values reported by national experts in the EUROMOD Country Reports. The corresponding numbers for 2019 are presented in Table A2 of Appendix 1.

4.2 The Selection of MI Beneficiaries

To apply this expenditure-based calibration, an iterative algorithm has been designed to select actual beneficiaries among the simulated eligible observations in each country until the total expenditure resulting from EUROMOD matches the total expenditure reported in the official administrative data. In our setting, the total simulated expenditure is defined as

{\overline{X}}_S

, whereas the total expenditure reported in the official source is

{\overline{X}}_O

. We aim to minimize an objective function

f(x)

by selecting a subset of actual beneficiary households

h

among all potential eligible households

H

such that

h\mathbf{\subseteq}H:\min \left(f(x)\right)

. Our objective function is consequently defined as:

f(x)=\left(1-\frac{{\overline{X}}_S}{{\overline{X}}_O}\cong 0.01\right)\ \mathrm{subject}\ \mathrm{to}\ {\overline{X}}_s>{\overline{X}}_O

To optimize the selection of the actual beneficiary households

h

, we follow Hernández, Picos, and Riscado (2022) and assume that the probability of receiving the benefit among eligible beneficiaries (

{P}_i

) depends on a deterministic and a random component.¹² Formally, for each household

i

{P}_i={w}^{\ast }{RC}_i+\left(1-w\right)\ast {DC}_i

where

{RC}_i\in \left(0,1\right)

is the random component, following a uniform distribution,

{DC}_i\in \left(0,1\right)

is the deterministic component, measuring the generosity of the entitlement, and

w

is the weight measuring the importance of each component in determining the probability.

{DC}_i

is given by:

{DC}_i=1-\frac{IT_i}{IT_i+{MI}_i}

where

{IT}_i

corresponds to the total income being subject to each MI scheme means-test, and

{MI}_i

is the minimum income benefit to which the household is entitled to. Consequently, the higher

{DC}_i

is, the more generous the entitlement, and the more likely the household is to be selected as an actual beneficiary.

Once ${P}_i$ has been computed for each household, an iterative algorithm is implemented in order to minimize our objective function. In practice, we define a calibration parameter $C=1$ that is sequentially decreased in 0.001 steps and confronted to ${P}_i$ . Then, for each subset of households $h$ whose ${P}_i>C$ , the resulting ${\overline{X}}_S$ is calculated and the objective function is solved until the latter is minimized for an optimal parameter ${C}^{\ast }$ . This process is repeated for each EU country.

Importantly, $w$ is treated as an exogenous parameter, and different values can be selected to balance the importance of the deterministic and random components. Under a full deterministic approach, such that $w=0$ , only the poorest households will end up being selected as actual beneficiaries.¹³ With a random component, such that $w>0$ , other elements beyond the generosity of the entitlement will weigh in the probability. This implies that households entitled to a low benefit might eventually end up being selected as beneficiaries. As argued in the studies on the drivers of non-take-up (Hernanz, Malherbet, and Pellizzari 2004), although the generosity of the entitlement plays a crucial role in determining whether to claim a benefit, other non-monetary factors are also relevant (e.g., individuals living in rural areas are less likely to take up the benefit). Our random component roughly approximates these non-monetary elements.

We parsimoniously assume that for all individuals the weights of the random and deterministic components are the same, that is, $w=0.5$ . Naturally, the results are sensitive to this choice. Since our algorithm selects beneficiaries as to align with a fixed level of expenditure, setting a higher weight for the random component would imply selecting more actual beneficiaries entitled to lower benefits, and vice versa. In Sub-section 5.4, we assess how changing the values of $w$ influences our main results.

4.3 Validation of the “Calibrated Baseline”

We present the results of the calibration in Table A2 in Appendix 1, while in Table A3 we provide some country-specific comments on the simulation and calibration of each scheme. To assess the accuracy of the calibration, we consider the ratio of simulated expenditure in EUROMOD to the actual expenditure, the “validation ratio.” We compute this indicator both for the “Full entitlement baseline,” and the “Calibrated baseline.” Expenditure in the “Full entitlement baseline” is overestimated by at least 30% for 18 countries. For many of them, the ratios are equal to two or even higher (e.g., Ireland [8], Estonia [3.5], Spain [3.2], and Bulgaria [3.1]). As expected, in the “Calibrated baseline” all ratios are very close to one. For countries with underestimation (ratios below 1) before the calibration procedure, the calibration has no effect, since it just selects all the eligible individuals without reaching the target expenditure. In some cases, ratios are close to one, which might indicate simultaneously accurate simulations and low non-take-up. Others are clearly lower than one (Cyprus, Czechia, and Germany¹⁴). As previously mentioned, it may be that some caveats of the simulations performed in EUROMOD are more prevalent in these countries, such as the inability to capture short-term entitlements.¹⁵ The consequences of MI overestimation are shown in Appendix 2, Figure A1, which illustrates that, for some countries, the “Full entitlement baseline” substantially underestimates the occurrence of zero and very low incomes as compared to EU-SILC. In the following sections, we use the “Calibrated baseline” as the benchmark for presenting the results.

5 Assessing the Effectiveness of EU MI Schemes in Poverty Alleviation

Having developed a corrected baseline simulation, we can turn to an evaluation of the effectiveness of existing MI schemes as simulated in this baseline. We perform this evaluation considering three main interconnected dimensions: the coverage and adequacy of the schemes, and their poverty-alleviation effects. We focus on working-age individuals, defined as those aged between 16 and 64, both inclusive. Throughout the whole paper, the poor population is defined as individuals in working-age whose equivalized disposable income is below the poverty threshold. This threshold is defined according to an extreme poverty criterion, calculated as 40% of the median equivalized disposable income of the whole population. In Appendix 3, Figures A2-A7 we also provide the main results according to a standard poverty criterion, using a 60% poverty line.

5.1 Coverage

Figure 2 presents estimates of the percentage of individuals benefitting, or not, from MI support. A beneficiary is defined as any individual living in a household for which at least one MI scheme has been granted. Results are presented in percentage of the total poor population of each country, prior to MI support, to ease comparability and take into account the fact that this is the main target population of the benefit. Individuals are classified into three different types according to their poverty status and receipt of MI support: (i) individuals that are poor before MI support and are beneficiaries (poor beneficiaries); (ii) individuals that are not poor before MI support but still receive it (non-poor beneficiaries); and (iii) individuals that are poor before MI support but are not beneficiaries (poor non-beneficiaries). In our setting, we define coverage as the share of poor beneficiaries, regardless of whether they remain in poverty or are lifted out of it after MI receipt, over the total poor population before MI (dark gray bar).¹⁶

MI schemes depict a heterogeneous coverage across EU Member States. Coverage rates vary from 6.5% in Latvia to almost 90% in France. Most countries fail in covering the majority of the poorest population, with only nine countries depicting coverage rates above 50%. The three countries showing the highest coverage are Cyprus, France, and Ireland, whereas the three countries with the lowest coverage are Bulgaria, Latvia, and Poland. In countries with the highest coverage, between 10% and 20% of the poorest individuals are left without MI support, whereas in countries with the lowest coverage this percentage exceeds 90%.

Coverage rates are influenced by the restrictiveness of the eligibility conditions and the extent of take-up. In countries with low coverage, very restrictive MI thresholds are typically in place, as detailed in the next subsection, limiting access to only the most impoverished individuals. Additionally, these countries do not generally apply income disregards for employment income in the eligibility conditions, which may further contribute to the low coverage rates. The existence of conditions related to owner-occupied housing may also reduce coverage. For instance, while Bulgaria has eligibility conditions related to owner-occupied housing,¹⁷ these are not present in Cyprus and Ireland. In France, the high coverage is likely associated with the inclusion of an employment bonus as part of the MI scheme.

Two key issues should be considered when assessing coverage. First, the targeting of MI schemes is imperfect in relation to the (monetary) poverty criteria used. As can be seen in Figure 2, a substantial number of individuals are identified as MI beneficiaries despite not being identified as poor (light gray bar), especially in countries that show high coverage rates, like France, Finland and Ireland. In some cases, the share of non-poor beneficiaries is as large, or even larger, as the share of poor non-beneficiaries, potentially suggesting a misallocation of MI support with respect to a strict (monetary) poverty criterion. This misalignment is largely due to MI thresholds being set above the 40% poverty line. When poverty is measured using a 60% poverty line, the extent of non-poor beneficiaries decreases (Figure A2 of Appendix 3).¹⁸

Second, broadening the definition of MI protection to include other schemes beyond our selection of “core” MI schemes would generally increase coverage in most countries, as shown in Table A7 of Appendix 5. This effect can be substantial in countries with very low coverage rates, such as Bulgaria, leading to some re-ranking within the EU. For instance, when housing benefits are taken into account, Bulgaria's coverage rate rises by more than 50 pp. However, even when accounting for all means-tested benefits, some countries, such as Italy, Latvia, Lithuania, and Poland, still show low coverage rates.

5.2 Adequacy

We now move to the assessment of the adequacy, that is, how generous MI schemes are. For this, we consider two approaches. The first approach is to use EUROMOD's Hypothetical Household Tool (HHoT), to assess adequacy for two “typical” household types, based solely on the rules of the tax-benefit system, without accounting for data-driven heterogeneities.¹⁹ The idea is to assess the level of adequacy across the EU in a comparative way, considering two households with the same characteristics in all countries. We select a single adult and a couple with two children as the two household types. Both units are jobless households, where all adults have been unemployed for the last 18 months, are actively seeking a job, and live in a rented accommodation.²⁰ This definition aims to approximate the maximum amount of the benefit that potential candidates to MI support could receive, triggering the simulation of the assessed MI schemes in EUROMOD.²¹ Figure 3 presents the resulting MI amounts as a share of each country-specific poverty line.

More than half of EU countries provide MI amounts that are not adequate with respect to an extreme poverty criterion, that is, the MI amount as a percentage of the 40% poverty line falls below 100%. In particular, the generosity of the schemes in Bulgaria, Czechia, Hungary, Latvia, Poland, and Romania is very low (below 40% for a single adult). Notably, in Bulgaria, Czechia, and Hungary, MI amounts are not adjusted regularly over time, but its indexation depends on political cycles and priorities, as well as the availability of resources (European Commission 2022a).²² On the contrary, Denmark, Finland, Germany, Ireland, Italy, Malta, and the Netherlands provide high levels of adequacy, above 120% for a single adult.

Looking at the differences between the two types of households, we find that in many countries the adequacy of MI support is lower for larger family units. With some exceptions, a couple with two children generally receives a less adequate benefit than a single adult. Two considerations are relevant here. The first relates to the use of the OECD-modified scale as an equivalence scale. According to this scale, a couple with two children aged below 14 years old should be able to spend 110% more than a single adult (50% for the partner, 30% for each child). Thus, to maintain the same adequacy level as of a single adult, MI amounts for a couple with two children should be at least 110% higher. However, our results suggest that the implicit equivalence scales of most EU MI schemes do not adequately address these additional needs, as measured by the OECD-modified scale. While this scale is widely accepted by researchers, it somewhat arbitrarily assigns values to the additional needs of family members. In reality, these needs vary according to a diverse variety of characteristics, including income (Garbuszus et al. 2021). More sophisticated approaches, such as those relying on expenditure data (e.g., Bargain and Donni 2012) or on the cost of goods and services required to meet basic needs (e.g., Penne et al. 2020), could be explored as alternatives.

A second consideration relates to complementary benefits. As previously discussed, it is often the case that alternative benefits complement MI provisions, thus increasing the adequacy (and coverage) of public cash support for low-income households. Child-related benefits, for example, are often provided in addition to MI benefits, with the former typically not being considered in the means-tests for the latter. As with coverage, we account for a broader range of schemes, beyond our identification of core MI schemes, and provide additional indicators of adequacy in Table A7 of Appendix 3. We find that the adequacy levels increase in several countries due to other complementary benefits, particularly child-related benefits, but that the increase remains limited in several countries (e.g., Bulgaria, Croatia, Hungary, and Romania) or is even zero in others (e.g., Italy, Spain).

The second approach we use to measure adequacy involves considering the specific characteristics of households receiving MI schemes in our calibrated data. In line with Figari, Matsaganis, and Sutherland (2013), we define adequacy as the share of poor individuals receiving MI schemes with equivalized disposable income above the poverty line. Figure 4 shows that in half of the countries, less than 50% of poor beneficiaries exit poverty after receiving the MI support. Only Austria, Denmark, Luxembourg, and the Netherlands show a high level of adequacy, above 90%. At the other extreme, in countries such as Estonia, Croatia, Latvia, and Romania, adequacy levels are very low, below 10%. These findings are consistent with those presented above, based on hypothetical households with predefined characteristics (Figure 3), and highlight that in most Central Eastern European countries and in some Southern European countries MI amounts are typically not adequate with respect to an extreme poverty criterion.

5.3 Poverty-Alleviation Effects

To the extent that countries offer a good combination of coverage and adequacy, this should translate into higher disposable incomes for the poorest individuals. Figure 5 depicts the mean annual equivalized disposable income of poor individuals before and after receiving MI support. Results account for the different price levels across EU Member States by dividing the nominal values (EUR) by the Price Level Index of each country as reported by Eurostat, to ease the comparability across countries. Averages are calculated for all poor individuals, independently of whether they receive MI support.

Before MI support, the equivalized disposable income of individuals in extreme poverty varies significantly across EU Member States, from 2000 to 6800 EUR per year (in real terms), reflecting the different socioeconomic conditions across EU countries before the last-resort safety nets took effect. The lowest values can be observed in Hungary, Greece, Romania, and Slovakia, while the highest ones correspond to Austria, Finland, France, and Luxembourg.

Remarkably, MI schemes do not enable the convergence of disposable incomes of the poorest citizens across the EU, as the best-performing countries before MI support are also those achieving the highest increases in disposable incomes due to MI support. In particular, the support provided by Belgium, France, Ireland, Luxembourg, and the Netherlands is particularly effective at improving the income conditions of the poorest citizens, whereas Bulgaria, Hungary, Latvia, and Romania's schemes barely entail any increase in average terms. This result aligns with the low coverage rates and adequacy levels found for the “core” MI schemes of these countries.

A complementary approach to evaluate the effectiveness of MI schemes in improving the income conditions of the poorest individuals is to look at poverty-related indicators. Figure 6 shows at-risk-of-poverty (AROP) rates before and after MI support for all EU countries. This indicator measures poverty incidence, that is, the share of the population with an equivalized disposable income below the poverty line. The poverty line is calculated based on the equivalized disposable income excluding MI support. Intuitively, we anchor the poverty line to the counterfactual scenario “before MI support” in which MI schemes are not in place.

In most EU countries, MI support seems insufficient to lift beneficiaries out of poverty, with reductions in AROP rates generally remaining below 1 pp. However, there are notable exceptions, including Cyprus (−2.4 pp), France (−2.5 pp), Germany (−2.5 pp), Ireland (−3.6 pp), and the Netherlands (−3.2 pp).²³

Figure 7 shows a second measure of poverty, AROP gaps, before and after MI support for all EU Member States. This indicator measures the intensity of poverty, that is, the mean shortfall in income from the poverty line, in percentage of the latter. Although existing EU MI schemes may not be sufficient to lift beneficiaries out of poverty, they do reduce the shortfall in income from the poverty line. Once again, there is significant heterogeneity across EU MI schemes, with a few countries reducing AROP gaps by more than 1 pp (e.g., Germany, Greece, Ireland, and the Netherlands), whereas others achieve only small reductions (e.g., Bulgaria, Czech Republic, Latvia, and Poland).

5.4 Evaluating Alternative Calibrations for Selecting MI Beneficiaries

As previously explained, in this paper we select actual MI beneficiaries from a pool of potential eligible units for MI support, aiming to meet a total simulated expenditure level that matches official expenditure data. The selection of actual MI beneficiaries can be carried out in various ways by combining different eligible units for MI support, thereby influencing our results. In this Sub-section, we present the sensitivity of our results—specifically coverage rates—to different calibration approaches. As outlined in Section 4, we define the probability of receiving a MI benefit based on a deterministic and a random component, as follows:

{P}_i={w}^{\ast }{RC}_i+\left(1-w\right)\ast {DC}_i

For the main results of this paper, we parsimoniously assign the same weight to the deterministic and the random components, setting w = 0.5. Here, we assess to what extent the value assigned to parameter $w$ influences our results, considering two additional calibrated scenarios (CS): (i) “Full random assignment” ( $w=1$ ), such that ${P}_i={RC}_i$ , scenario CS1; (ii) “Full deterministic assignment” $\left(w=0\right)$ , with higher probability of receiving a benefit for lower income households, such that ${P}_i={DC}_i$ , scenario CS2. The literature on the determinants of the take-up rate (e.g., Van Oorschot 1991) suggests that the probability of claiming a benefit positively correlates with the size of the benefit (i.e., w < 1, and P = ${w}^{\ast }{RC}_i+\left(1-w\right)\ast {DC}_i\Big)$ . Hence, we would expect that the real scenario lies somewhere between CS1 and CS2.

Table 2 shows differences in coverage rates with respect to the baseline. First, we present the “Full entitlement baseline” to underscore that this scenario substantially overestimates coverage in most EU countries. Recall that this scenario informs on the potential impact of MI support under full take-up, although it might be influenced by the absence or partial simulation in EUROMOD of certain eligibility conditions to access MI support. The largest differences between our default calibration (w = 0.5) and the “Full entitlement baseline” are observed in Austria, Luxembourg, Slovakia, and Spain. As explained in Section 3, these differences might be attributed both to the absence or partial simulation of certain rules to access MI support²⁴, and high non-take-up rates. For instance, in Luxembourg, due to a lack of information, wealth-related requirements are not simulated in EUROMOD. Similarly, in Austria, the asset test is only partly simulated. In Spain, conditions such as job-seeker registration, application for other benefits before seeking MI support, or not owning secondary residences cannot be simulated with precision.²⁵ Additionally, existing studies indicate substantial non-take-up rates in these countries (Amétépé 2012; Eurofound 2015; Fuchs et al. 2020).

TABLE 2. Differences in coverage rates under various calibration approaches.

Country	Baseline (w = 0.5)	“Full entitlement baseline” (no calibration)	(w = 1)	(w = 0)
		Difference w.r.t the baseline
			Calibrated scenarios
			CS1: “Full random assignment”	CS2: “Full deterministic assignment”
AT	23	75	11	−6
BE	50	42	−14	9
BG	7	22	2	−1
CY	84	0	0	0
CZ	47	0	0	0
DE	65	0	0	0
DK	36	13	−7	5
EE	12	43	5	0
EL	64	0	0	0
ES	27	47	10	−1
FI	71	26	−9	17
FR	89	1	−7	0
HR	27	1	1	−2
HU	11	4	0	4
IE	80	9	−6	−1
IT	18	26	3	−1
LT	24	6	0	−1
LU	41	45	0	−7
LV	6	15	3	−1
MT	41	21	6	0
NL	53	0	0	0
PL	10	27	6	0
PT	50	12	8	−6
RO	32	9	−1	1
SE	38	23	4	−3
SI	76	18	1	8
SK	43	48	−2	−4

Note: Difference in percentages points with respect to our default calibration scenario (w = 0.5).
Source: Our own elaboration using EUROMOD.

Next, we examine the impact of the alternative calibrated scenarios, CS1 and CS2. Compared to our baseline calibration (w = 0.5), CS2 typically involves selecting fewer but poorer beneficiaries, leading to lower coverage rates, while CS1 produces coverage rates that are generally similar to or higher than the baseline calibration. These findings reflect the influence of the deterministic component, ${DC}_i$ , in determining the probability to be selected as actual MI beneficiary. However, there are some exceptions. In countries like Finland, Denmark or Belgium, our default calibration (w = 0.5) includes many non-poor beneficiaries (see Figure 2), whereas CS2 tends to select poorer beneficiaries. As coverage is calculated as a proportion of poor individuals, this results in higher coverage rates under CS2 than in the baseline for these countries. Conversely, for CS1, we observe the opposite: in Belgium, Finland, Denmark, and other few countries, some poor beneficiaries included in our default calibration may be excluded under a full random assignment, leading to lower coverage than in the baseline. Despite these exceptions, CS1 and CS2 do not show large differences across most EU countries when compared to our default calibration.

Similar comparisons can be made based on additional indicators. In Appendix 2, Tables A5 and A6, we present the results on adequacy and poverty, respectively. Regarding adequacy, recall that we measure it through the share of poor individuals receiving MI schemes whose equivalized disposable income rises above the poverty line. As this indicator is limited to MI beneficiaries in poverty, it is not affected by the possible selection of non-poor individuals. While it shows some variability especially depending on other characteristics beyond income, such as household type (e.g., single or couple with children) or the level of assets owned, it is not that sensitive to the changes between different calibrated scenarios.²⁶ Regarding poverty, CS2 tends to produce lower AROP gaps compared to the baseline, because of the selection of MI beneficiaries who typically exhibit higher poverty gaps and are therefore entitled to higher MI amounts. On the contrary, CS1 generally results in higher AROP gaps compared to the baseline. The variation in AROP rates is mostly stable, and we do not observe substantial discrepancies with respect to the default calibration. Changes in AROP rates across different calibrations largely depend on how transitions around the poverty threshold take place. In countries where the adequacy of the MI schemes is not sufficient to lift beneficiaries out of poverty, calibrating via the use of different values of our parameter w does not entail significant changes in AROP rates, as the equivalized disposable income of MI beneficiaries remains below the poverty threshold.

6 Exploring Possible Reforms to Existing MI Schemes

In light of the low effectiveness of existing MI schemes in poverty alleviation in several EU countries, we explore the simulation of possible reforms. We consider changes to the two main elements defining the performance of MI schemes: coverage and adequacy. The reforms considered are theoretical, focused on the working-age population, and not targeted at changing country-specific aspects of the MI scheme of each country. The goal of the simulations is not to provide a comprehensive analysis of all the possible reforms that could be considered by authorities in each country but rather to: (i) provide a methodological framework to think about the main aspects that should be considered when reforming MI schemes; (ii) show how far/close the current MI scheme for each EU country is from eradicating extreme poverty; and (iii) show how this would change if adequacy (first) and coverage (second) improved.

6.1 The Reform Scenarios: Description and Rationale

We simulate a new hypothetical complementary MI scheme, which works as a last-resort safety net for the poorest households, complementing their equivalized disposable incomes up to each country-specific extreme poverty line. The eligibility to the new complementary scheme is only made on a purely monetary basis, with no additional criteria being considered (i.e., there are no age-related criteria, wealth tests, or other requirements potentially excluding those in extreme monetary poverty from being eligible for the complement). The unit of assessment is the household, and thus the incomes of all cohabiting individuals are pooled together regardless of their family links. This differs with respect to most MI schemes in EU Member States, where the unit of assessment usually uses a narrower concept (i.e., cohabiting individuals linked by family relations up to a specific degree).²⁷ The scheme operates after the simulation of all taxes and benefits, including each existing country-specific MI scheme(s). Hence, the final benefit amount is not taxable, nor included in other means-tested benefits. The amount of the complementary MI support for each household is calculated as the difference between households' equivalized disposable income and each country-specific extreme poverty line.²⁸

Once the new scheme is simulated, we restrict its accessibility to three different populations of interest by considering three possible reform scenarios, following a stepwise approach. First, the new complementary MI scheme is only assigned to the current beneficiaries of each country's existing MI scheme. Therefore, if the current MI scheme is not adequate to bring the corresponding beneficiary household out of extreme poverty, the new MI scheme financially complements this household's disposable income up to the 40% poverty line. Beneficiaries already receiving an adequate amount are not complemented further. This scenario (“Increased adequacy” scenario, hereinafter) shows the impact of increasing the generosity of the MI entitlement for the current beneficiaries in poverty by the amount needed to lift them out of extreme poverty. Second, and in addition to the previous scenario, we assign the new complementary MI scheme to a specific number of individuals in extreme poverty who were not receiving the existing country-specific MI scheme such that the coverage rate of MI support increases by 10 pp.²⁹ This scenario (“Increased coverage” scenario, hereinafter) shows the effects of reaching a higher proportion of those in need by increasing the number of beneficiaries of MI support.³⁰ Third, the new complementary MI scheme is provided to all poor households that had not been selected as beneficiaries in the previous two scenarios. This final scenario (“Extreme poverty elimination” scenario, hereinafter) measures the impact of eliminating extreme poverty in each country and provides a benchmark to which the current MI scheme and the previously simulated scenarios can be compared to.

The results from these simulations are then compared with the ones for the “Calibrated baseline,” as presented in Section 5.³¹ Importantly, our results should be interpreted as “morning-after effects,” as we abstract from potential behavioral responses triggered by the simulated reforms. These responses might distort the outcomes of a specific policy change, moving away from the original objectives. Regarding MI schemes, behavioral responses usually refer to labor supply disincentives (see, for instance, Bargain and Doorley 2011), justifying the introduction of accompanying labor activation strategies (e.g., participation in training programs, tailored assistance with job vacancies). Furthermore, our results do not capture potential consumption and saving behavioral reactions that might occur as a response to changes in households' disposable incomes (see, among others, Nelson 2012, on the relationship between minimum income levels and material deprivation).

6.2 What Improvements in the Living Standards of Poor Households?

We now focus on households whose income falls below the extreme poverty line, despite the existing MI schemes. Figure 8 shows the percentage change in the mean annual equivalized disposable income of individuals in extreme poverty implied by each reform scenario relative to the baseline.

The average equivalized disposable income of individuals in extreme poverty for the EU27 would increase by 4.1% in the “Increased adequacy” scenario, with significant variation across EU Member States. In some countries like Hungary, Lithuania, Portugal, and Romania, an increase in adequacy would imply a non-negligible increase in household disposable income, between 15% and 25%. This is not surprising, since these countries were among those performing worst in terms of adequacy. In contrast, in countries like Austria, Denmark, Luxembourg, and the Netherlands, which already had a high adequacy in the baseline scenario, an increase in adequacy would barely rise disposable income.

When considering an increase in both adequacy and coverage, that is, the “Increased coverage” scenario, the EU27 average equivalized disposable income of individuals in extreme poverty would increase by 9.4%, again with significant heterogeneity. In countries like Hungary, Lithuania, Portugal, and Romania, the increase would be between 23% and 33%, suggesting that in these countries there might be room for enhancing the targeting of these schemes via the promotion of the take-up and/or the adjustment of the eligibility criteria toward the inclusion of those in monetary extreme poverty. On the contrary, in countries like Finland, Ireland, and Slovenia, the increase would be below 4%, suggesting that the potential of reforms to existing MI schemes in increasing poor households' disposable income is somewhat limited.

Finally, the average equivalized disposable income of EU27 poor individuals would increase by 20.1% in the “Poverty elimination” scenario, reflecting both the change in the level of the benefit for current beneficiaries and the access to the new adequate entitlement for previously excluded individuals. As before, there would be significant differences across countries, with some countries such as Hungary, Lithuania, and Romania having increases around 60%, and other countries like Cyprus, Ireland and Finland having increases below 6%.

Figure 9 shows the AROP gaps in the baseline and in each reform scenario. The relative impacts of increasing adequacy and coverage differ substantially across countries. In some countries, such as Bulgaria, Lithuania, and Italy, an increase in adequacy to the poverty line would lead to an only negligible decrease in the degree of poverty intensity. An increase in coverage by 10 pp., however, would lead to a substantial decrease in the AROP gap. In other countries, such as Greece, Romania, and Slovakia, the opposite is true, with an increase in adequacy being more effective at reducing the distance between the income of poor individuals to the poverty line. Taken together, these two hypothetical reforms substantially impact the AROP gap in most countries, decreasing its level at the EU27 average by 0.7 pp., with five countries (Greece, Hungary, Italy, Lithuania and Romania) achieving reductions exceeding 1 pp. Increasing coverage to target poverty elimination would bring the AROP gaps close to zero in all countries.³² For most countries, a substantial decrease in the AROP gap would be achieved in this third scenario, suggesting that efforts toward reaching excluded households from MI support would be the key element of potential reforms.

6.3 How Far From Extreme Poverty Eradication?

Table 3 shows the total expenditure on existing MI schemes, as a percentage of GDP, as well as the expenditure changes implied by our reform scenarios, in pp with respect to the “Calibrated baseline.”

TABLE 3. Expenditure on existing Minimum Income (MI) schemes and on simulated reform scenarios.

Country	“Calibrated baseline” (% GDP)	“Increased adequacy” scenario	“Increased coverage” scenario	“Extreme poverty elimination” scenario
		Diff. w.r.t baseline (pp)
AT	0.23	+0.00	+0.06	+0.17
BE	0.30	+0.00	+0.04	+0.07
BG	0.03	+0.03	+0.10	+0.35
CY	0.61	+0.03	+0.04	+0.05
CZ	0.03	+0.02	+0.03	+0.05
DE	0.54	+0.03	+0.10	+0.15
DK	0.53	+0.01	+0.05	+0.17
EE	0.06	+0.04	+0.11	+0.29
EL	0.30	+0.10	+0.14	+0.16
ES	0.27	+0.05	+0.13	+0.34
FI	0.31	+0.01	+0.01	+0.02
FR	0.78	+0.03	+0.05	+0.05
HR	0.10	+0.07	+0.12	+0.27
HU	0.06	+0.04	+0.12	+0.39
IE	0.45	+0.01	+0.02	+0.03
IT	0.28	+0.02	+0.14	+0.51
LT	0.13	+0.04	+0.13	+0.35
LU	0.24	+0.00	+0.03	+0.10
LV	0.02	+0.03	+0.12	+0.39
MT	0.19	+0.02	+0.05	+0.08
NL	0.60	+0.00	+0.06	+0.17
PL	0.02	+0.04	+0.08	+0.19
PT	0.15	+0.08	+0.10	+0.15
RO	0.05	+0.13	+0.19	+0.35
SE	0.23	+0.01	+0.08	+0.20
SI	0.55	+0.01	+0.02	+0.03
SK	0.15	+0.05	+0.09	+0.16
EU27	0.43	+0.03	+0.09	+0.20

Note: EU27 stands for the GDP weighted EU average.
Source: Our own elaboration using EUROMOD and Eurostat data on GDP (2019).

At the EU27 average level, moving to the “Increased adequacy” scenario entails an increase in the expenditure to GDP ratio of ~0.03 pp., reflecting the higher amounts of MI support implied by moving all MI beneficiaries to each country-specific extreme poverty line. An increase in expenditure is observed in all countries, but the values of these increases are highly heterogeneous. They are particularly high in Croatia, Greece, Portugal, and Romania suggesting that the provision of more generous amounts to the current beneficiary households in those countries could have significant budgetary consequences.

Moving to the “Increased coverage” scenario, we observe an increase in the EU27 expenditure to GDP ratio of ~0.10 pp. relative to the baseline scenario. The increase in coverage rates is, therefore, more costly than the increase in adequacy levels. At the country level, the increases in expenditure are once again highly heterogeneous, varying from 0.01 pp. in Finland, where the initial coverage rate is among the highest in the EU, up to 0.19 pp. in Romania, whose initial coverage is less well targeted.³³

Finally, considering the “Poverty elimination” scenario, the expenditure to GDP ratio for the EU27 average would increase from 0.43% to 0.63%. In other words, to achieve “Poverty Elimination” the increase in expenditures would be of 0.20 pp. relative to the baseline, illustrating that the additional cost of providing MI support to lift all poor households in the EU out of poverty would be low.³⁴ For some countries, such as Cyprus, Finland, France, Ireland, and Slovenia, the additional expenditure would be rather small, suggesting that poverty elimination would require a relatively marginal effort. For others, such as Bulgaria, Hungary, Italy, Latvia, Lithuania, Spain, and Romania, the increase in expenditure would vary between 0.30 pp. and 0.55 pp., indicating that complete poverty elimination could be a somewhat more ambitious objective given the existing coverage and adequacy levels.

As previously mentioned, our estimates must be interpreted as pure “morning-after” effects, with two important caveats applying: first, as shown by Collado et al. (2016), if work incentives are to be considered such that the existing participation tax rates remain unchanged, the cost of lifting individuals out of poverty would be significantly higher. In this sense, our estimates only represent lower bounds of the real costs these (theoretical) reforms would have. Second, and even in the presence of negative labor supply responses, an extended provision of MI support could also have some self-financing properties if currently social excluded individuals are eventually reintegrated into the labor market, for example, by means of the activation inclusion elements usually embedded within these schemes (Marchal and van Mechelen 2017). Hence, a comprehensive evaluation of the fiscal viability of eradicating extreme poverty should ideally consider these two elements and adopt a more dynamic approach, including macroeconomic feedback effects (Barrios et al. 2019).

7 Conclusion and Discussion

MI schemes are an important tool available to policymakers to alleviate poverty and fight social exclusion, by guaranteeing a last-resort safety net for households with insufficient resources. An accurate assessment of the effectiveness of these schemes in terms of their design and reach is crucial to identify potential gaps and avenues for reforms. Despite its relevance, this assessment faces several data and modeling limitations, which often lead to an under or overestimation of the extent of MI support, and to a bias in the measurement of the real impact of this support on disposable incomes and poverty.

In this paper, we attempt to provide an integrated and consistent evaluation of the effectiveness of MI schemes in poverty alleviation in the EU in three main steps. First, we apply a new method to calibrate the simulation of MI schemes in the microsimulation model EUROMOD, to tackle the limitations of the model and obtain a new “closer to real spending” baseline simulation of each EU Member State's scheme(s). Second, we use this corrected baseline to evaluate existing MI schemes, investigating their degree of coverage and adequacy, their impacts on poverty, and their overall cost. Third, we explore the effects of possible (theoretical) reforms, implementing sequential changes to the levels of coverage and adequacy, and calculate the “morning-after” budgetary cost of extreme poverty elimination.

Our results suggest that the coverage rate of existing MI schemes is quite heterogeneous across EU countries but generally insufficient, with most countries reaching less than 50% of households in extreme poverty and some having coverage rates even below 10%. Intuitively, this might explain why during the COVID-19 crisis several countries felt the need to implement temporary schemes for social assistance and other discretionary measures (see e.g., Christl et al. 2024), somewhat calling into doubt the potential automatic stabilization properties of existing MI schemes. Moreover, we find that the benefit levels provided by existing schemes are not adequate to provide a minimum standard of living in more than half of the EU countries, with the generosity of the support being below 50% of the extreme poverty line in several countries. Furthermore, the adequacy of MI support is often lower for larger families vis-à-vis single-person households. This suggests that the complementary amounts provided for children through MI support may be insufficient to adequately cover the expenses of raising them. While many EU countries ensure adequate income levels for families with children by providing additional child-related benefits alongside MI support, in some countries the combined support remains insufficient. The combination of the different coverage and adequacy levels across EU countries results in heterogeneous effects on poverty alleviation. Remarkably, MI schemes do not enable the convergence of disposable incomes of the poorest citizens across the EU, as the best-performing countries before MI support are also those with the highest increases in disposable incomes through MI support. In addition, MI support in most EU countries seems insufficient to lift beneficiaries out of poverty. Nevertheless, it does play an important role in reducing the shortfall in incomes from the poverty line.

There is scope for overcoming some of the gaps in current MI schemes through reforms affecting both the coverage and adequacy of these schemes. Although with high heterogeneity across EU countries, our results suggest that extending the coverage to all poor households in working-age, such that extreme poverty is fully eliminated, would imply an overall additional cost of 0.20 pp. relative to the expenditure to GDP ratio produced by existing MI schemes, at the EU level. Therefore, the additional “morning-after” cost of providing MI support to lift all poor households in the EU out of poverty (relative to the status quo) would be rather low, and arguably far from unattainable.

The analysis performed in this paper provides a framework to think about the main aspects to consider when evaluating and reforming MI schemes and gives fresh insights into the effectiveness of current MI schemes in the EU and possible ways to reform them. The main takeaways are that the contribution of MI support to poverty elimination is still rather limited in some EU countries and that action could be taken to increase coverage and adequacy, moving toward poverty elimination, at a relatively low budgetary cost. Implementing these changes can be achieved through different means: enabling better accessibility to the MI schemes (e.g., promoting take-up by improving access to information, decreasing bureaucratic procedures, fighting stigmatization), softening the eligibility criteria (e.g., perhaps reconsidering the role of asset tests³⁵ and other non-income conditions, such as minimum age), moving toward a closer consideration of adequacy vis-à-vis poverty thresholds, and so forth. Our analysis also discusses the importance of the definition of MI schemes when assessing their poverty effects. While in some countries coverage rates increase significantly when considering a broader definition of MI protection, in other countries these rates are almost unaffected.

Despite its potential usefulness, the analysis is, naturally, not without limitations. First, our results are pure “morning-after” effects, abstracting from any possible behavioral reactions to the simulated reforms. Indeed, our approach only sets up a static benchmark to assess the current performance of MI schemes and how this would change under policy reforms aimed at increasing their adequacy and coverage. A more comprehensive assessment should incorporate a dynamic approach, with a concrete focus on behavioral reactions, both from supply (e.g., labor supply effects) and demand (e.g., changes in consumption and savings) perspectives. Second, the reforms considered are theoretical and do not account for any country-specific aspects of the design of MI schemes, simply exploring the effects of changing coverage and adequacy in a uniform way for all countries. A more complete and informative analysis should provide a discussion and evaluation of the different reform possibilities available for each country based on the characteristics of the country's scheme. Finally, the results are sensitive to the approach adopted for the calibration of the new baseline. A more accurate analysis should perform more refined calibrations considering the availability of more disaggregated data at country-specific levels.

The above-discussed limitations provide important avenues for future research. In particular, considering behavioral effects on work incentives and consumption reactions to MI reforms would be an important step toward a more comprehensive analysis of the effectiveness of MI schemes. In the same vein, the simulations could be extended to include not only cash transfers but also in-kind benefits, such as free childcare, which in some countries may play an important role in ensuring a minimum living standard.

Acknowledgments

The views expressed are purely those of the authors and may not in any circumstances be regarded as stating an official position of the European Commission or the Organisation for Economic Co-operation and Development. The results presented here are based on EUROMOD version I3.86+. Originally maintained, developed, and managed by the Institute for Social and Economic Research (ISER), since 2021, EUROMOD has been maintained, developed, and managed by the Joint Research Centre (JRC) of the European Commission in collaboration with Eurostat and national teams from the EU countries. We are indebted to many people who have contributed to the development of EUROMOD. The results and their interpretation are the authors' responsibility. We acknowledge the fruitful comments of Ana Agúndez, Anna Burger, Olivier Bontout and Viginta Ivaškaitė-Tamošiūnė, and of the participants of the XXIX Meeting on Public Economics, the EUROMOD Annual Meeting and Research Workshop 2022, the OECD Applied Economics Work-in-Progress Seminar, and the 10th Meeting of the Society for the Study of Economic Inequality (ECINEQ). Comments from anonymous referees are also gratefully acknowledged. We immensely thank Fidel Picos for his valuable insights and support with the methodology.

Conflicts of Interest

The authors declare no conflicts of interest.

Endnotes

¹ For instance, in the EU Statistics on Income and Living Conditions (SILC) microdata, MI schemes are often reported jointly with other “social exclusion benefits not elsewhere classified”, impeding an accurate assessment of their impacts in isolation from other cash benefits (Trindade and Goedemé 2020).

² We focus on MI schemes defined at the national level, that is, those whose policy rules are the same in the whole country. In cases where there is large regional variation in MI rules, such as Spain, these schemes are also accounted for. However, we do not consider MI policies defined at the local level, that is, those available only in specific municipalities. Although these policies may play a significant role in reducing poverty and social exclusion, they are not modeled in EUROMOD. For country-specific details, see Table A3 in Appendix 1.

³ Except for Germany, Ireland and Malta, where unemployment assistance schemes can rather be assimilated to MI schemes, as confirmed by national experts.

⁴ The list of MI schemes was consulted with EU and national experts in the context of the European Commission Proposal for a Council Recommendation on adequate Minimum Income ensuring active inclusion.

⁵ An exception is Spain, where we also simulate a scheme introduced in 2020 to account for the importance of this reform in MI support.

⁶ Table A1 in Appendix 1 compares the list of MI schemes as reported in MISSOC, based on the schemes in force in 2019, and our selection of MI schemes for this paper. The schemes listed in MISSOC and not selected in this paper are either targeted only to the elderly, designed for special circumstances, or considered as unemployment assistance.

⁷ We specifically use the EUROMOD version I3.86+. For a detailed description of EUROMOD, see Sutherland and Figari, Matsaganis, and Sutherland (2013) and https://euromod-web.jrc.ec.europa.eu/ (last accessed September 1, 2024).

⁸ Although 2020–2021 policy years were already available in version I3.86+ of EUROMOD, we decided to use 2019 as the main policy year for two main reasons. First, in our analysis, we aim to capture the main structural elements of MI schemes, leaving aside the exceptional circumstances of the COVID-19 crisis in terms of labor market conditions and temporary policy responses. Second, the availability of administrative official data on expenditure in MI schemes comes with delay, while our calibration crucially relies on these data. Given that official statistics might not be available for the latest policy years, we use the 2019 system as a way of reconciling the simulated and official figures. The only exceptions pertain to Latvia and Spain, where we also simulate the structural changes introduced to their schemes in 2020 to acknowledge the importance of these structural changes in MI support. In Italy, although the scheme was introduced in April 2019, we assume it was in effect for the entire year to ensure comparability with the other countries.

⁹ With the exceptions in the latter case of Belgium, where a take-up correction is performed.

¹⁰ Investment incomes are present in the EU-SILC data, and they are used in some countries to derive financial assets (by capitalizing those incomes). However, this information is usually insufficient to account for the full extent of the assets' portfolio.

¹¹ This categorization is in line with Hernanz, Malherbet, and Pellizzari (2004), which provides an extensive literature review on the extent and drivers of non-take-up across OECD countries.

¹² Differently from Hernández, Picos, and Riscado (2022), we do not calibrate the model based jointly on the number of beneficiaries and expenditure, but we focus only on the latter. Note that several approaches can be implemented to select the actual beneficiaries. In this paper, we opt for a rather simple strategy, which is applied in a harmonized way to all EU Member States. However, in countries where, for instance, administrative microdata on beneficiaries of MI schemes are available, the selection of beneficiaries could be performed using more refined methodologies (e.g., estimating the likelihood to be an actual beneficiary based on the socioeconomic characteristics of the eligible population).

¹³ Typically, the lower the household income is, the higher the benefit. Hence, under the deterministic approach, it is more likely that households with very low income will be selected.

¹⁴ Besides potential caveats of the simulation, the underestimation of MI schemes in Germany could be driven by the small sample size (less than 0.3% of the population), which results in limited coverage of certain population subgroups (Hauser 2008; Frick and Krell 2010).

¹⁵ Moreover, it is worth noting that expenditure statistics may also suffer from some comparability issues (e.g., some residual schemes being accounted for in the administrative figures but not in the simulated results).

¹⁶ This is broadly in line with the concept used by Figari, Matsaganis, and Sutherland (2013). Note that the concept of coverage has been subject to considerable discussion (see Nelson and Nieuwenhuis 2021 for a recent contribution toward consolidating a framework for its analysis).

¹⁷ Note that Bulgaria is among the EU countries with the highest share of people living in households owning their home (i.e., 85%).

¹⁸ Other factors that may also contribute to this discrepancy are the use of different equivalence scales in MI rules as compared with the OECD modified scale, and/or narrower definitions of the MI assessment unit vis-à-vis the household.

¹⁹ The Hypothetical Household Tool (HHoT) is a EUROMOD tool used for constructing hypothetical households and generating data according to a set of chosen household and individual characteristics. It generates a hypothetical dataset, which can then be used to estimate the effects of tax-benefit policies on household disposable income in a “controlled” environment, according to a well-defined set of characteristics. The range of possible characteristics is large, including labor market status, types of income sources, gender, and age, among others. For more information, see Hufkens et al. (2019).

²⁰ The housing costs are assumed to be 20% of the country-specific average monthly gross earnings.

²¹ Naturally, a different definition of households' characteristics might lead to different levels of adequacy.

²² According to the European Commission (2022a), other countries where MI amounts are not adjusted regularly are Greece, Italy, Portugal and Croatia. Most of these countries also show low levels of adequacy. Nevertheless, as highlighted by Van Mechelen and Marchal (2013), simply relying on regular indexation may not be sufficient to prevent the erosion of social assistance benefits. Ad-hoc reforms might be needed to address this issue effectively over time.

²³ Note that unemployment assistance in Ireland and Germany (Jobseeker's Allowance and Grundsicherung für Arbeitsuchende, respectively), and a benefit paid to low-income earners (Prime d'activité) in France, are included in the list of MI schemes.

²⁴ Note that the absence or partial simulation of MI rules is present in both the “Full entitlement baseline” and the calibrated scenarios. However, in the latter, by selecting only certain units from the pool of eligible households, we effectively exclude some units from MI support, as if they were not meeting certain conditions that we are unable to simulate.

²⁵ The limitations in the simulation of MI schemes in EUROMOD for each country are reported in detail in the corresponding EUROMOD Country Reports.

²⁶ Recall that in Section 5 we also measured adequacy using hypothetical families with predefined characteristics (Figure 3). In this case, the calibration is not applicable and thus our results remain unchanged.

²⁷ The use of the household as the assessment unit maintains consistency with the standard measurement of incomes in distributional and poverty analyses, which equalizes household disposable incomes to account for differences in household size and composition.

²⁸ As shown by the results in Appendix 5, MI schemes in the EU do not seem designed to address standard poverty (i.e., considering a 60% poverty line) and are far from achieving this purpose. Therefore, the hypothetical scenarios in this section focus on a less costly and ambitious objective, which is eradicating extreme poverty (i.e., considering a 40% poverty line).

²⁹ The highest coverage rates (in the baseline) are close to 90%; hence, in these countries, an increase by 10 pp. would already lead to full coverage. In order to maintain cross-country comparability, we decided to apply the same increase in coverage rate (10 pp) in all countries.

³⁰ We do not specify how this increase in the number of beneficiaries could be achieved. Two options could be to either enhance the take-up of the schemes or relax some eligibility criteria (e.g., related to age or assets ownership).

³¹ The impacts of the three reform scenarios on coverage, adequacy, and AROP rates are determined by construction, and therefore we do not present them in the main text. However, it is useful to briefly describe these impacts to render the mechanics of the reform scenarios, which we do in Appendix 4.

³² Income losses are not compensated by MI schemes, so households with net negative incomes still remain below the poverty line, which explains why the AROP gaps are not exactly zero.

³³ Noticeably, the cost of this scenario would depend on how far from the 40% poverty line are the new beneficiaries that make up the additional 10 pp. increase in the coverage rate.

³⁴ To provide an idea of scale, it is worth noting that total expenditure on social protection amounts to 28% of GDP in the EU27, based on Eurostat data (last accessed 1 September 2024): https://ec.europa.eu/eurostat/databrowser/view/tps00098/default/table?lang=en&category=t_spr.

³⁵ See Marchal et al. (2021) for a comprehensive analysis of the impact of asset tests within MI schemes.

³⁶ For a few countries, namely Greece, Poland and Slovakia, the coverage rate does not exactly reach 100%. This can be explained by the larger extent of negative disposable incomes in these countries, which usually arise in presence of self-employed individuals that might have to pay fixed social insurance contributions regardless of their level of incomes. Individuals with negative disposable incomes are treated by the new complementary MI scheme as if they had zero incomes so that the top-up benefit does not go above the guaranteed MI level.

³⁷ As previously mentioned, income losses are not compensated by the MI schemes, so a few households with net negative incomes still remain below the poverty line.

³⁸ Note here that in our definition of coverage the denominator corresponds to the total poor population before MI support, whereas some of the additional benefits reported in Table A7 might not be targeted to the poor, especially those benefits without means-tested characteristics.

Appendix 1: List of MI Schemes and Validation Results

TABLE A1. Classification of guaranteed MI resources according to MISSOC and selection of schemes for this paper.

Country	MI scheme as in MISSOC	EUROMOD policy	Target group
AT	Guaranteed minimum resources (Mindestsicherung)	bsa_at	Working-age and elderly
BE	Integration income (Revenu d'intégration/leefloon)	bsa_be	Working-age
	Guarantee of income for elder persons (garantie de revenus aux personnes âgées/inkomensgarantie voor ouderen)		Elderly
	Income Replacement Allowance (allocation de remplacement de revenus/inkomensvervangende tegemoetkoming)		Special case (disability)
	Integration Allowance (allocation d'intégration/integratietegemoetkoming)		Special case (disability)
	Allowance for Assistance to the Elderly (allocation pour l'aide aux personnes âgées/tegemoetkoming voor hulp aan bejaarden)		Special case (elderly care)
BG	Social assistance allowances (Месечни социални помощи)	bsa00_bg	Working-age and elderly
BG	Social pension for old-age (Социална пенсия за старост)		Elderly
CY	Guaranteed Minimum Income (Eλάχιστo Eγγυημένo Eισόδημα)	bsamm_cy	Working-age and elderly
	Social Pension (Koινωνική ∑ $\overset{\acute{\mkern6mu}}{\upsilon }$ νταξη).		Elderly
	Scheme supporting pensioners' households with low income (∑χέδιo ενίσχυσης νoικoκυρι $\overset{\acute{\mkern6mu}}{\upomega}$ ν συνταξιo $\overset{\acute{\mkern6mu}}{\upsilon }$ χων με χαμηλά εισoδήματα)		Elderly
CZ	Allowance for Living (Příspěvek na živobytí)	bsa_cz	Working-age and elderly
	Supplement for Housing (Doplatek na bydlení)		Special case (housing)
	Extraordinary Immediate Assistance (Mimořádná okamžitá pomoc)		Special case (extraordinary circumstances, e.g., natural disaster)
DE	Subsistence benefit (Hilfe zum Lebensunterhalt)	bsa00_de	Working-age and elderly
	Unemployment assistance for jobseekers (Grundsicherung für Arbeitsuchende)	bunnc_de	Working-age
	Assistance for the elderly and for persons with reduced earning capacity (Grundsicherung im Alter und bei Erwerbsminderung)		Elderly and special case (disability)
DK	Social assistance (kontanthjælp)	bsa_dk	Working-age
	Integration benefits (integrationsydelse)		Special case (integration of non-residents, e.g., refugees)
	Educational assistance (uddannelseshjælp)		Special case (students)
EE	Subsistence benefit (toimetulekutoetus)	bsa00_ee	Working-age and elderly
	Unemployment allowance (töötutoetus)		Working-age
	Allowance for old-age pensioner living alone (üksi elava pensionäri toetus)		Elderly
EL	Social Solidarity Income (“SSI”) (KOINΩNIKO EI∑OΔHMΑ ΑΛΛHΛEΓΓYH∑)	bsa00_el	Working-age and elderly
EL	Social solidarity allowance for uninsured elders (EΠIΔOMΑ KOINΩNIKH∑ ΑΛΛHΛEΓΓYH∑ ΑNΑ∑ΦΑΛI∑TΩN YΠEPHΛIKΩN)		Elderly
ES	Regional Minimum Income Schemes (Rentas Mínimas de Inserción)	bsarg_es	Working-age
	Minimum Living Income (Ingreso Minimo Vital)	bsa00_es	Working-age
	Non-contributory old-age benefit (Pensión de jubilación no contributiva)		Elderly
	Non-contributory invalidity benefit (Pensión de invalidez no contributiva)		Special case (disability)
	Unemployment assistance (desempleo de nivel asistencial)		Working-age
	Minimum for Spanish persons residing abroad and returnees (Prestación por razón de necesidad a favor de los españoles residentes en el exterior y retornados)		Special case (integration of non-residents, e.g., refugees)
FI	Basic Social Assistance (Perustoimeentulotuki)	bsa00_fi	Working-age and elderly
FR	Active solidarity income (revenu de solidarité active, RSA)	bsa00_fr	Working-age
	Employment bonus (Prime d'activité)	bsawk_fr	Working-age
	Allowance for disabled adults (allocation pour adulte handicapé, AAH)		Special case (disability)
	Solidarity allowance for the elderly (allocation de solidarité aux personnes âgées, ASPA)		Elderly
	Supplementary disability allowance (allocation supplémentaire d'invalidité, ASI)		Special case (disability)
	Allowance of specific solidarity (allocation de solidarité spécifique, ASS)		Working-age
HR	Guaranteed minimum benefit (Zajamčena minimalna naknada)	bsa_hr	Working-age and elderly
HU	Benefit for persons in active age (aktív korúak ellátása)	bsa_hu	Working-age
HU	Old-age allowance (Időskorúak járadéka)		Elderly
IE	Supplementary Welfare Allowance	bsa00_ie	Working-age and elderly
	Jobseeker's Allowance	bunnc_ie	Working-age
	Disability Allowance		Special case (disability)
	Blind Pension		Special case (disability)
	Carer's Allowance		Special case (care)
	Farm Assist		Special case (farmers)
	Widow's, Widower's or Surviving Civil Partner's		Special case (widowhood)
	State Pension		Elderly
IT	Guaranteed Minimum Income or Pension (Reddito o Pensione di Cittadinanza)	bsamm_it	Working-age and elderly
	Social allowance (Assegno sociale)		Elderly
	Guaranteed Minimum Pension (Pensione di Cittadinanza)		Elderly
LT	Cash social assistance (Piniginė socialinė parama)	bsa00_lt	Working-age and elderly
LT	Social assistance pensions (Šalpos pensija)		Elderly and special case (disability)
LU	Social inclusion income (revenu d'inclusion sociale, Revis)	bsacm_lu	Working-age and elderly
LV	Guaranteed minimum income (GMI) benefit (Pabalsts garantētā minimālā ienākuma līmeņa nodrošināšanai)	bsamm_lv	Working-age and elderly
MT	Social assistance (Ghajnuna Socjali)	bsa_mt	Working-age
MT	Unemployment Assistance (Gℏajnuna gℏal-Diżimpjieg)	bunncmt_mt	Working-age
NL	Participation Act (Participatiewet)	bsagross_nl	Working-age and elderly
NL	Participation Act (Participatiewet)	bsanet_nl	Working-age and elderly
PL	Periodic Allowance (Zasiłek okresowy)	ben_sa_pl	Working-age and elderly
PL	Permanent Allowance (Zasiłek Stały):		Elderly and special case (disability)
PT	Social integration minimum income (Rendimento social de inserção)	bsa00_pt	Working-age and elderly
	Old-age social pension (pensão social de velhice)		Elderly
	Widow(er)'s pension (pensão de viuvez)		Special case (widowhood)
	Orphan's pension (pensão de orfandade)		Special case (orphanhood)
	Social Inclusion Benefit (Prestação Social para a Inclusão)		Special case (disability)
	Solidarity supplement for the elderly (complemento solidário para pessoas idosas)		Elderly
	Social allowances in the framework of maternity, paternity and adoption protection (subsídios sociais de protecção na parentalidade)		Special case (parental leave)
	Unemployment allowance (subsídio social de desemprego)		Working age
RO	Social Aid (Ajutor social)	bsa_ro	Working-age and elderly
RO	Heating allowances		Special case (housing)
SE	Social assistance (Ekonomiskt bistånd)	bsamt_se	Working-age and elderly
SI	Financial Social Assistance (denarna socialna pomoč)	bsa_si	Working-age and elderly
SI	Supplementary Allowance (varstveni dodatek)		Elderly and special case (disability)
SK	Material Need Benefit (Dávka v hmotnej núdzi)	bsa_sk	Working-age and elderly

Note: The selected MI schemes for this paper are highlighted in gray. The schemes listed are based on MISSOC data as of July 2019, except for Spain, where we also include the nationwide MI scheme implemented in early 2020.
Source: MISSOC and EUROMOD Country Reports.

TABLE A2. Validation of minimum income (MI) schemes, before and after calibration.

Country	EUROMOD policy	X (1)	N (2)	X (3)	N (4)	X (5)	N (6)	X (1)/(5)	N (2)/(6)	X (3)/(5)	N (4)/(6)
		Full entitlement baseline		Calibrated scenario		Official statistics		Validation—“Full entitlement baseline”		Validation—“Calibrated scenario”
AT	bsa_at	2583.5	430.7	913.6	69.9	913.0	154.0	2.8	2.8	1.0	0.5
BE	bsa_be	3163.8	504.1	1441.4	157.4	1453.3	208.5	2.2	2.4	1.0	0.8
BG	bsa00_bg	113.0	92.4	36.9	32.1	36.0	32.0	3.1	2.9	1.0	1.0
CY	bsamm_cy	141.3	31.5	141.3	31.5	199.0	43.7	0.7	0.7	0.7	0.7
CZ	bsa_cz	1875.1	68.4	1875.1	68.4	2750.2	63.0	0.7	1.1	0.7	1.1
DE	bunnc_de	17946.7	2781.9	17946.7	2781.9	27924.0	2747.0	0.6	1.0	0.6	1.0
DE	bsa00_de	748.3	191.2	748.3	191.2	1500.0	113.0	0.5	1.7	0.5	1.7
DK	bsa_dk	23717.3	221.8	12288.2	97.1	12383.0	205.0	1.9	1.1	1.0	0.5
EE	bsa00_ee	55.5	28.3	16.4	6.1	16.0	10.9	3.5	2.6	1.0	0.6
EL	bsa00_el	554.6	248.0	554.6	248.0	665.0	265.5	0.8	0.9	0.8	0.9
ES	bsa00_es	4589.2	1125.9	1390.8	208.5	1400.0	260.2	3.3	4.3	1.0	0.8
ES	bsarg_es	1933.2	717.1	1929.5	379.7	1970.2	369.3	1.0	1.9	1.0	1.0
FI	bsa00_fi	1110.6	298.7	746.3	174.0	753.0	297.0	1.5	1.0	1.0	0.6
FR	bsa00_fr	11873.0	3743.2	10103.8	2914.1	10200.0	1667.5	1.2	2.2	1.0	1.7
FR	bsawk_fr	9442.4	4298.8	8920.1	3908.1	9004.0	4201.0	1.0	1.0	1.0	0.9
HR	bsa_hr	421.6	45.0	416.9	42.5	421.0	35.1	1.0	1.3	1.0	1.2
HU	bsa_hu	44094.7	159.7	28494.0	92.2	28618.0	135.0	1.5	1.2	1.0	0.7
IE	bunnc_ie	1501.1	230.8	1501.1	230.8	1629.0	139.0	0.9	1.7	0.9	1.7
IE	bsa00_ie	959.5	126.4	120.3	7.4	120.0	16.0	8.0	7.9	1.0	0.5
IT	bsamm_it	9628.2	1591.3	5063.1	589.2	5100.0	1114.0	1.9	1.4	1.0	0.5
LT	bsa00_lt	88.2	56.8	64.0	34.7	63.4	64.6	1.4	0.9	1.0	0.5
LU	bsacm_lu	326.9	27.6	150.1	8.6	151.0	9.7	2.2	2.8	1.0	0.9
LV	bsamm_lv	13.0	29.0	4.9	7.7	4.7	17.2	2.8	1.7	1.0	0.5
MT	bunmt_mt	4.3	1.8	4.0	1.0	3.9	1.1	1.1	1.6	1.0	0.9
MT	bsa_mt	38.5	12.6	22.8	5.6	22.4	5.0	1.7	2.5	1.0	1.1
NL	bsagross_nl, bsanet_nl	4904.2	553.4	4904.2	553.4	5659.0	459.0	0.9	1.2	0.9	1.2
PL	ben_sa_pl	1287.3	377.2	557.9	105.6	552.0	285.0	2.3	1.3	1.0	0.4
PT	bsa00_pt	348.9	144.9	327.1	117.5	329.9	128.0	1.1	1.1	1.0	0.9
RO	bsa_ro	775.2	469.9	564.5	325.1	559.4	176.0	1.4	2.7	1.0	1.8
SE	bsamt_se	17055.0	357.4	11528.1	171.6	11594.0	248.4	1.5	1.4	1.0	0.7
SI	bsa_si	343.5	99.3	267.8	60.0	269.0	101.0	1.3	1.0	1.0	0.6
SK	bsa_sk	316.5	172.2	137.1	44.0	135.0	63.0	2.3	2.7	1.0	0.7

Note: Simulations correspond to 2019. For LV and ES, the permanent changes introduced to MI schemes in 2020 have been included. Expenditure results are expressed in MIL., national currency. Number of beneficiaries in EUROMOD corresponds to heads of the different MI assessment units and are expressed in thousands.
Source: Our own elaboration using EUROMOD. Official statistics are obtained from EUROMOD Country Reports, and they correspond to 2019 (latest available data in most countries) unless otherwise stated in the column “comments” of Table A3.

TABLE A3. Country-specific comments on the simulation and calibration of minimum income (MI) schemes in EUROMOD.

Country	EUROMOD policy	MI scheme	Comments
AT	bsa_at	Guaranteed minimum resources (Mindestsicherung)	MI benefit rules are partly different in all federal states. Only the system of Vienna is simulated in EUROMOD
BE	bsa_be	Integration income (Revenu d'intégration/leefloon)	Original non-take-up adjustment in place in the public version of EUROMOD for the MI scheme has been replaced by the new calibration
BG	bsa00_bg	Monthly social assistance allowances (Mесечни социални помощи)
CY	bsamm_cy	Guaranteed MI (Eλάχιστo Eγγυημένo Eισόδημα)	Number of official beneficiaries not available in EUROMOD Country Reports and obtained from OECD—Social Benefits Recipients Database, 2016
CZ	bsa_cz	Allowance for Living (Příspěvek na živobytí)
DE	bsa00_de	Subsistence benefit (Hilfe zum Lebensunterhalt)
DE	bunnc_de	Unemployment assistance for jobseekers (Grundsicherung für Arbeitsuchende)
DK	bsa_dk	Social assistance (Kontanthjælp)
EE	bsa00_ee	Subsistence benefit (Toimetulekutoetus)	Original non-take-up adjustment in place in the public version of EUROMOD for the MIS has been replaced by the new calibration
EL	bsa00_el	Guaranteed MI (EΛΑXI∑TΟ EΓΓYHMENΟ EI∑ΟΔHMΑ)
ES	bsarg_es	Regional MI Schemes (Rentas Mínimas de Inserción)	Special calibration is being applied to preserve the heterogeneous coverage of MI schemes across regions. Official data corresponding to 2020 (to be consistent with the enter in force of the national scheme)
ES	bsa00_es	Minimum Living Income (Ingreso Minimo Vital)	Official data obtained from national sources, corresponding to May 2021: converted to annual terms assuming the same beneficiaries/expenditure over the year
FI	bsa00_fi	Social assistance (Toimeentulotuki)	Original non-take-up adjustment in place in the public version of EUROMOD for the MI has been replaced by the new calibration
FR	bsa00_fr	Active solidarity income (Revenu de solidarité active, RSA)	Original non-take-up adjustment in place in the public version of EUROMOD for the MI has been replaced by the new calibration
FR	bsawk_fr	Employment bonus (Prime d'activité)
HR	bsa_hr	Guaranteed minimum benefit (Zajamčena minimalna naknada)	Original non-take-up adjustment in place in the public version of EUROMOD for the MI has been replaced by the new calibration
HU	bsa_hu	Benefit for persons in active age (Aktív korúak ellátása)
IE	bsa00_ie	Supplementary Welfare Allowance	Original non-take-up adjustment in place in the public version of EUROMOD for the MI has been replaced by the new calibration
IE	bunnc_ie	Jobseeker's Allowance
IT	bsamm_it	Guaranteed MI (Reddito di Cittadinanza)	Preliminary official data obtained from national sources, corresponding to 2019. The Reddito di Cittadinanza was introduced in April, but we assume it was in place over the whole year for consistency with other countries
LT	bsa00_lt	Social benefit (Socialinė pašalpa)
LU	bsacm_lu	Social inclusion income (Revenu d'inclusion sociale, Revis)
LV	bsamm_lv	Guaranteed MI benefit (Pabalsts garantētā minimālā ienākuma līmeņa nodrošināšanai)	MI is strictly connected to the housing benefit. We assume that if the calibration sets the MI for a household to 0, the housing benefit is also set to 0 (same take-up behavior for both benefits) Municipality differences cannot be accounted for in the model, as there is no detailed information on people's residences. Only the rules of Riga municipality are simulated (as this is the biggest municipality)
MT	bsa_mt	Social assistance (Ghajnuna Socjali)
MT	bunmt_mt	Unemployment Assistance (Gℏajnuna gℏal-Diżimpjieg)
NL	bsagross_nl bsanet_nl	Participation Act (Participatiewet)
PL	ben_sa_pl	Periodic Allowance (Zasiłek okresowy)	Original non-take-up adjustment in place in the public version of EUROMOD for the MI has been replaced by the new calibration
PT	bsa00_pt	Social MI (Rendimento social de inserção)
RO	bsa_ro	Social Aid (Ajutor social)	Original non-take-up adjustment in place in the public version of EUROMOD for the MI has been replaced by the new calibration
SE	bsamt_se	Social assistance–livelihood support (Ekonomiskt bistånd)
SI	bsa_si	Financial Social Assistance (Denarna socialna pomoč)	Original non-take-up adjustment in place in the public version of EUROMOD for the MI has been replaced by the new calibration
SK	bsa_sk	Material Need Assistance (Pomoc v hmotnej núdzi)	EUROMOD simulations include also additional supplementary allowances for: housing, protection, activation and dependent children. Disentangling all these allowances from the main one was not feasible and seemed inconsistent with the official expenditure data Original non-take-up adjustment in place in the public version of EUROMOD for the MI has been replaced by the new calibration

Source: Our own elaboration using EUROMOD Country Reports.

TABLE A4. Main rules of EU minimum income (MI) schemes, 2019.

EU country		Qualifying residence period	Age	Owner-occupied housing	Other immovable and movable assets	Job search	Employment income disregard	Provision of complementary allowances	Taxation
		Non-income-related					Income-related
				Assets
Austria	AT	5+ years for third country nationals	No age conditions	Yes	Yes (with some exceptions)	Yes, otherwise decrease of value or even suspension of benefit	Amount/share of earnings partially exempted	Housing, healthcare, and heating allowances can be granted together with the MI scheme	Not taxable
Belgium	BE	3+ months	18–65	No	Yes	Yes, otherwise benefit can be denied	Amount/share of earnings partially exempted	Housing (region-specific), healthcare, and heating allowances can be granted together with the MI scheme	Not taxable
Bulgaria	BG	5+ years for third country nationals	No age conditions	Yes	Yes	Yes, otherwise benefit can be withdrawn	No disregard, with minor exceptions (e.g., earned income from programs promoting employment)	Housing, healthcare, and heating allowances can be granted together with the MI scheme	Not taxable
Cyprus	CY	5+ years	28+ (with exceptions: e.g., single parents)	No	Yes	Yes, otherwise benefit can be withdrawn	Amount/share of earnings partially exempted	Housing (region-specific) and healthcare allowances can be granted together with the MI scheme	Not taxable
Czechia	CZ	No minimum period required	No age conditions	No	Yes	Yes, otherwise benefit can be temporarily suspended	Amount/share of earnings partially exempted	Housing allowances can be granted together with the MI scheme	Not taxable
Germany	DE	3+ months	No age conditions for subsistence benefit; 15+ for unemployment assistance	Yes if the house considered not to be “appropriate”	Yes (with some exceptions)	Yes	Amount/share of earnings partially exempted	Housing and heating allowances can be granted together with the MI scheme	Not taxable
Denmark	DK	9+ years of residence in the previous 10 years	From 30 to retirement age	Yes	Yes	Yes, otherwise benefit can be withdrawn	Amount/share of earnings partially exempted	Housing and healthcare allowances can be granted together with the MI scheme	Taxable
Estonia	EE	No minimum period required	No age conditions	Yes	Yes	Yes, otherwise decrease of value or even suspension of benefit	Amount/share of earnings partially exempted	No separate allowances	Not taxable
Greece	EL	No minimum period required	18+	Yes	Yes	Yes	Amount/share of earnings partially exempted	No separate allowances	Not taxable
Spain	ES	1+ year immediately before application	23+	No	Yes	Yes	No disregard	Housing allowances can be granted together with the MI scheme	Taxable
Finland	FI	No minimum period required	No age conditions	No	Yes	Yes, otherwise decrease of value of the benefit	Amount/share of earnings partially exempted	Housing and healthcare allowances can be granted together with the MI scheme	Not taxable
France	FR	5+ years for third country nationals	25+ and 18+ for young workers in some cases	Yes	Yes	Yes	Amount/share of earnings partially exempted	Housing allowances can be granted together with the MI scheme	Not taxable
Croatia	HR	No minimum period required	No age conditions	Yes	Yes	Yes, otherwise benefit can be temporarily suspended	Amount/share of earnings partially exempted	Housing and healthcare allowances can be granted together with the MI scheme	Not taxable
Hungary	HU	No minimum period required	18+ up to retirement age	No	Yes	Yes, otherwise benefit can be withdrawn	Amount/share of earnings partially exempted	Housing and transport allowances (decided at the local level) can be granted together with the MI scheme	Not taxable
Ireland	IE	No minimum period required	No age conditions for supplementary welfare allowance; 18–65 for jobseeker's allowance	No	Yes	Yes, otherwise decrease in value or temporary suspension of benefit	No disregard	Housing, healthcare, and fuel allowances can be granted together with the MI scheme	Not taxable
Italy	IT	10+ years, 2+ consecutive	No age conditions	No	Yes	Yes, otherwise benefit can be withdrawn	No disregard in general, with some exceptions	Housing and healthcare allowances can be granted together with the MI scheme	Not taxable
Lithuania	LT	No minimum period required	No age conditions	Yes	Yes	Yes, otherwise benefit can be suspended or denied	Amount/share of earnings partially exempted	Energy allowance can be granted together with the MI scheme	Not taxable
Luxembourg	LU	5+ years during last 20 years for third-country nationals	25+ (with some exceptions)	Yes	Yes	Yes, otherwise decrease of value or temporary suspension of benefit	Amount/share of earnings partially exempted	Rent allowance can be granted together with the MI scheme	Taxable
Latvia	LV	No minimum period required	No age conditions	No	Yes	Yes, otherwise decrease in value of the benefit	No disregard	Housing, healthcare, and transport allowances (decided at the local level) as well as in-kind energy allowances can be granted together with the MI scheme	Not taxable
Malta	MT	No minimum period required	18–60	No	Yes	Yes, otherwise benefit can be suspended	No disregard	Rent, healthcare, and energy allowances can be granted together with the MI scheme	Not taxable
Netherlands	NL	No minimum period required	18+	Yes	Yes	Yes, otherwise decrease of value or temporary suspension of benefit	No disregard in general, with some exceptions	Rent allowance can be granted together with the MI scheme	Taxable
Poland	PL	No minimum period required	18+	No	No (unless too high)	May be required, but subject to discretionary judgment	No disregard	Housing allowance can be granted together with the MI scheme	Not taxable
Portugal	PT	1+ year for third-country nationals	18+ (with some exceptions)	Partly, if exceeds an index	Yes	Yes	Amount/share of earnings partially exempted	Healthcare fee exemptions can be granted together with the MI scheme	Not taxable
Romania	RO	No minimum period required	18+	No	Yes	Yes, otherwise benefit can be suspended or withdrawn	No disregard	Heating allowances can be granted together with the MI scheme	Not taxable
Sweden	SE	No minimum period required	No age conditions	Yes	Yes	Yes, otherwise decrease in value or temporary suspension of benefit	Amount/share of earnings partially exempted	Housing allowance and coverage of some health expenditures can be granted together with the MI scheme	Not taxable
Slovenia	SI	No minimum period required	No age conditions	No	Yes	Yes, otherwise benefit can be denied	No disregard	Housing allowances and healthcare fee exemptions can be granted together with the MI scheme	Not taxable
Slovakia	SK	No minimum period required	18+	No	Yes	Yes, otherwise benefit can be suspended	Amount/share of earnings partially exempted	Housing allowance can be granted together with the MI scheme	Not taxable

Source: Our own elaboration based on EUROMOD Country Reports, MISSOC database and European Commission (2022a, 2022b).

Appendix 2: Additional Results of the Calibration

Figure A1 compares the income distribution in the “Full entitlement baseline,” in the EU-SILC data, and in the “Calibrated baseline” for countries with a “validation ratio” over 2. The distribution is depicted only for incomes below 60% of each country-specific median equivalized disposable income. Generally, the “Full entitlement baseline” shifts the income distribution to the right compared to the EU-SILC data, with this shift being particularly noticeable in Austria, Spain, Luxembourg, and Slovakia. Notably, income frequencies for zero or low-income values are barely observable in the “Full entitlement baseline” due to the model oversimulation of MI support, thereby influencing poverty-related calculations. For instance, in Austria, the “Full entitlement baseline” produces no individuals placed below 40% of the median equivalized disposable income. In other countries, the impact of the overestimation might not be as pronounced, yet we still typically observe a lower frequency of zero and low-income values with respect to the EU-SILC data. The “Calibrated baseline” generally better resembles the income distribution observed in EU-SILC.

TABLE A5. Differences in adequacy levels under various calibration approaches.

Country	Baseline (w = 0.5)	“Full entitlement baseline” (no calibration)	CS1: “Full random assignment” (w = 1)	CS2: “Full deterministic assignment” (w = 0)
		Difference w.r.t the baseline
			Calibrated scenarios
AT	97	−2	-5	−1
BE	88	−5	−4	0
BG	14	3	5	9
CY	69	0	0	0
CZ	30	0	0	0
DE	57	0	0	0
DK	92	2	−2	1
EE	2	18	3	−1
EL	23	0	0	0
ES	33	−7	−8	4
FI	82	−2	−4	4
FR	61	0	0	1
HR	8	0	0	1
HU	14	10	4	10
IE	83	5	0	0
IT	59	−4	−2	1
LT	26	0	−3	2
LU	97	0	2	−1
LV	8	−5	−1	−5
MT	39	−2	−6	0
NL	90	0	0	0
PL	19	−5	−6	−6
PT	17	−1	0	2
RO	5	1	3	2
SE	60	−1	−1	1
SI	73	0	1	3
SK	12	14	16	−6

Note: Differences in percentage points with respect to our default calibration scenario (w = 0.5).
Source: Our own elaboration using EUROMOD.

TABLE A6. Differences in AROP rates and gaps under various calibration approaches.

Country	Before MI	(no calibration)	(w = 1)	(w = 0)	(w = 0.5)	Before MI	(no calibration)	(w = 1)	(w = 0)	(w = 0.5)
	AROP rates					AROP gaps
		Difference w.r.t before MI					Difference w.r.t before MI
			Calibrated scenarios					Calibrated scenarios
		“Full entitlement baseline”	CS1: “Full random”	CS2: “Full deterministic”	Baseline		“Full entitlement baseline”	CS1: “Full random”	CS2: “Full deterministic”	Baseline
AT	5.8	−5.4	−1.8	−0.9	−1.3	2.0	−2.0	−0.7	−0.8	−0.8
BE	3.6	−2.8	−1.1	−1.8	−1.6	1.7	−1.6	−0.6	−1.4	−1.1
BG	9.1	−0.4	−0.2	−0.1	−0.1	2.9	−0.4	−0.1	−0.1	−0.1
CY	4.3	−2.4	−2.4	−2.4	−2.4	1.5	−1.2	−1.2	−1.2	−1.2
CZ	2.4	−0.3	−0.3	−0.3	−0.3	0.7	−0.2	−0.2	−0.2	−0.2
DE	6.7	−2.5	−2.5	−2.5	−2.5	3.1	−2.0	−2.0	−2.0	−2.0
DK	6.4	−3.0	−1.7	−2.4	−2.1	2.7	−1.8	−1.2	−1.6	−1.5
EE	6.7	−0.7	−0.1	0.0	0.0	2.6	−1.1	−0.3	−0.4	−0.4
EL	8.2	−1.2	−1.2	−1.2	−1.2	3.5	−1.7	−1.7	−1.7	−1.7
ES	10.3	−2.0	−1.0	−1.0	−0.9	4.2	−2.8	−1.6	−1.5	−1.5
FI	3.0	−2.3	−1.5	−2.3	−1.8	0.8	−0.7	−0.5	−0.7	−0.6
FR	4.5	−2.5	−2.3	−2.5	−2.5	1.7	−1.3	−1.2	−1.3	−1.3
HR	6.9	−0.2	−0.2	−0.2	−0.2	2.7	−0.6	−0.6	−0.6	−0.6
HU	12.0	−0.4	−0.2	−0.4	−0.2	5.5	−0.3	−0.2	−0.3	−0.3
IE	5.5	−4.3	−3.4	−3.6	−3.6	2.1	−1.9	−1.6	−1.8	−1.8
IT	11.1	−2.7	−1.3	−1.1	−1.1	5.5	−2.3	−1.1	−1.4	−1.2
LT	7.6	−0.6	−0.4	−0.5	−0.5	3.7	−0.7	−0.6	−0.7	−0.7
LU	5.0	−4.2	−2.0	−1.6	−2.0	1.9	−1.6	−0.8	−1.0	−1.0
LV	8.3	−0.1	−0.1	0.0	0.0	3.5	−0.6	−0.2	−0.2	−0.2
MT	3.6	−0.9	−0.6	−0.6	−0.6	1.6	−0.9	−0.8	−0.8	−0.8
NL	6.7	−3.2	−3.2	−3.2	−3.2	3.1	−1.8	−1.8	−1.8	−1.8
PL	5.6	−0.3	−0.1	−0.1	−0.1	2.1	−0.3	−0.1	−0.2	−0.2
PT	6.7	−0.7	−0.7	−0.6	−0.6	2.7	−1.3	−1.2	−1.2	−1.2
RO	11.1	−0.3	−0.3	−0.3	−0.2	4.7	−0.5	−0.4	−0.5	−0.5
SE	7.0	−2.5	−1.8	−1.5	−1.6	3.0	−1.7	−1.2	−1.5	−1.4
SI	3.8	−2.6	−2.1	−2.4	−2.1	1.5	−1.4	−1.2	−1.4	−1.3
SK	5.3	−1.2	−0.6	−0.1	−0.3	2.8	−2.0	−0.9	−1.3	−1.3

Note: Difference in percentages points with respect to scenario before MI support, with poverty lines anchored to this scenario.
Source: Our own elaboration using EUROMOD.

Appendix 3: Results of the Effectiveness of Existing Schemes Using the 60% Poverty Line

Appendix 4: Additional Results of the Reform Scenarios

Figure A8 shows the coverage rate of MI schemes in the “Calibrated baseline” and the three reform scenarios. For all countries, moving from the “Calibrated baseline” to the “Increased adequacy” scenario entails no impact in coverage, as it only implies a change in the amounts received by each current beneficiary, without generating any new recipient. Therefore, the coverage rates remain the same as those obtained in Figure 2. The amounts provided become perfectly adequate, as they allow all beneficiaries to be lifted out of extreme poverty.

Moving from the “Increased adequacy” scenario to the “Increased coverage” scenario implies an increase in the coverage rate of 10 pp. Noticeably, this translates into very different absolute increases in the number of beneficiaries, depending on the initial coverage rate. For countries with very low initial coverage rates (e.g., Poland, Estonia, Bulgaria, and Latvia), the number of beneficiaries more than doubles, while for countries with high initial coverage rates (e.g., France, Cyprus, Ireland, and Slovenia), the change is relatively small. Adequacy levels remain the same as in the previous scenario, as all the beneficiaries, “old” and “new,” are assumed to receive the necessary amount to reach the 40% poverty line.

Finally, moving from the “Increased coverage” scenario to the “Poverty elimination” scenario effectively implies increasing the coverage rate up to 100% for almost all countries.³⁶ This is expected given that in this scenario all individuals who are in extreme poverty receive MI support such that their equivalized disposable income equals the 40% poverty line. As in the previous scenarios, there is full adequacy.

Figure A9 shows the AROP rates in the “Calibrated baseline” and the three reform scenarios. We find that the “Increased adequacy” scenario would lead to a non-negligible decrease in extreme poverty in Romania, Greece, and Portugal. This confirms our previous findings, indicating that an increase in adequacy in these countries would be effective at increasing households' disposable income and, hence, reducing the risk of extreme poverty. At the EU27 average level, the AROP rate would decrease from 5% to 4.1%. Implementing the “Increased coverage” scenario (i.e., enlarging each current country-specific coverage rate by 10 pp) would lead to an additional reduction in the AROP rate in all EU countries, in particular in Bulgaria, Italy, and Spain, although in most countries the reduction would not be substantial and would be smaller than the reduction obtained in the first scenario. Lastly, as expected, the “Poverty elimination” scenario would bring the AROP rates close to zero in all countries.³⁷ For most countries, a substantial decrease in the AROP rate would be achieved in this third scenario, suggesting that efforts toward reaching potentially excluded households from MI support would be the key element of potential reforms.

Appendix 5: Sensitivity of Coverage and Adequacy to Different Definitions of MI Protection

Table A7 shows a comparison of the coverage of MI schemes under the “core” definition used in this study with the coverage using broader definitions, including additional schemes to which poor individuals might be entitled. As expected, the inclusion of additional benefits, such as other social assistance schemes and housing benefits, increases the coverage with respect to our initial measurement. The increases are particularly large in Bulgaria, Denmark, Luxembourg, Malta, and Romania, whereas other countries remain almost unaffected (e.g., Belgium, Italy, or Spain). Accounting for all means-tested benefits, or eventually all types of benefits (including non-means-tested schemes), enlarge further the coverage rates, yet the protection provided by countries like Croatia, Estonia, Hungary, Latvia, Lithuania, or Italy still remains under the lowest across the EU, no matter the measure of coverage being considered.³⁸

TABLE A7. Coverage of Minimum Income (MI) protection (%), according to different definitions of MI support.

Country	“Core” MI schemes	“Core” MI schemes + other social assistance and housing benefits	All means- tested benefits	All benefits
AT	22.8	29.6	62.3	78.3
BE	49.8	50.8	71.9	82.4
BG	7.1	57.5	82.6	85.4
CY	83.5	91.0	91.2	93.2
CZ	46.9	79.7	90.7	90.7
DE	64.9	71.2	73.3	82.1
DK	36.4	100.0	100.0	100.0
EE	11.6	16.8	13.0	69.4
EL	63.9	74.9	83.1	84.9
ES	26.8	29.4	73.3	78.4
FI	71.0	97.3	99.4	99.4
FR	89.1	95.9	97.4	97.4
HR	27.4	32.7	55.3	64.7
HU	11.1	23.2	57.3	64.4
IE	79.8	91.1	91.3	96.2
IT	17.5	19.6	41.2	53.9
LT	24.0	43.7	55.3	59.5
LU	41.1	99.3	99.3	99.3
LV	6.5	29.0	25.0	60.8
MT	41.1	84.6	95.4	95.4
NL	52.9	74.2	98.5	99.1
PL	10.4	20.9	38.7	77.9
PT	50.1	53.0	71.1	73.2
RO	32.1	80.0	89.6	90.2
SE	37.9	71.7	68.5	83.9
SI	76.2	80.1	84.6	89.9
SK	42.8	48.4	60.2	88.8

Note: In this case, results are only shown for the extreme poverty criterion (40% of median equivalized disposable income). Importantly, our calibration applies only to the MI schemes, with other additional benefits simulated as in the publicly available EUROMOD baseline. The latter might result in overestimated coverage rates if these additional benefits are also oversimulated.
Source: Our own elaboration using EUROMOD.

In addition to the adequacy provided by each country-specific “core” MI schemes, Table A8 reports adequacy levels based on the final disposable income (i.e., after all benefits and taxes have been simulated) of the same hypothetical households considered in Section 5. As it can be seen, the addition of other social benefits (e.g., housing benefits) increases notably the adequacy of MI protection for a single individual in some countries (e.g., Czechia, France, Latvia, or Malta), it remains unchanged in others (e.g., Austria, Belgium, Italy, or Spain) or it even decreases in some cases (e.g., Denmark or the Netherlands), the latter having to do with the schemes of those countries being taxable. More generally, higher levels of adequacy can be now observed for a couple with two children across most EU countries, with some exceptions (e.g., Italy, Spain), given the entitlement of this sort of hypothetical household to additional child-related benefits.

TABLE A8. Adequacy of Minimum Income (MI) protection (% of extreme poverty line) by household types, including other social benefits.

Country	MI scheme	Disposable income	MI scheme	Disposable income
	Single		Couple with two children
AT	114.7	114.7	94.6	122.5
BE	111.6	111.6	73.3	104.9
BG	18.1	28.5	19.7	42.0
CY	112.3	112.3	100.7	108.8
CZ	30.3	105.4	45.7	85.8
DE	155.6	155.6	96.8	133.7
DK	149.6	126.4	189.3	157.1
EE	40.8	40.8	64.8	79.1
EL	66.0	89.1	40.9	90.4
ES	90.3	90.3	81.7	81.7
FI	132.3	141.5	104.2	123.4
FR	64.8	105.0	55.2	97.2
HR	41.7	41.7	39.7	54.6
HU	34.9	34.9	35.5	55.6
IE	159.8	159.8	153.2	169.3
IT	140.1	140.1	100.9	100.9
LT	43.8	43.8	66.7	90.8
LU	116.7	116.4	90.2	116.1
LV	23.0	62.5	49.2	85.3
MT	127.3	177.4	149.2	171.2
NL	153.1	119.1	93.6	102.2
PL	30.8	30.8	42.3	102.0
PT	54.1	54.1	69.6	87.1
RO	21.1	21.1	20.8	53.8
SE	122.0	122.0	72.0	109.2
SI	86.4	86.4	92.8	117.9
SK	65.7	65.7	71.4	79.6

Note: In this case, results are only shown for the extreme poverty criterion (40% of median equivalized disposable income).
Source: Our own elaboration using EUROMOD.

References

Amétépé, F. 2012. “The Effectiveness of Luxembourg's Minimum Guaranteed Income.” International Social Security Review 65: 99–116. https://doi.org/10.1111/j.1468-246X.2011.01421.x.
10.1111/j.1468-246X.2011.01421.x
Google Scholar
Ayala, L., J. M. Arranz, C. García-Serrano, and L. Martínez-Virto. 2021. “The Effectiveness of Minimum Income Benefits in Poverty Reduction in Spain.” International Journal of Social Welfare 30: 152–169. https://doi.org/10.1111/ijsw.12447.
10.1111/ijsw.12447
Web of Science® Google Scholar
Ayala, L., and E. Bárcena-Martín. 2020. “Measuring Social Welfare Gains in Social Assistance Programs: An Application to European Countries.” Social Indicators Research 151: 205–229. https://doi.org/10.1007/s11205-020-02377-w.
10.1007/s11205-020-02377-w
Web of Science® Google Scholar
Bárcena-Martín, E., M. C. Blanco-Arana, and S. Pérez-Moreno. 2018. “Social Transfers and Child Poverty in European Countries: Pro-Poor Targeting or pro-Child Targeting?” Journal of Social Policy 47: 739–758. https://doi.org/10.1017/S0047279418000090.
10.1017/S0047279418000090
Google Scholar
Bargain, O., and O. Donni. 2012. “Expenditure on Children: A Rothbarth-Type Method Consistent With Scale Economies and Parents' Bargaining.” European Economic Review 56: 792–813. https://doi.org/10.1016/j.euroecorev.2012.02.003.
10.1016/j.euroecorev.2012.02.003
Web of Science® Google Scholar
Bargain, O., and K. Doorley. 2011. “Caught in the Trap? Welfare's Disincentive and the Labor Supply of Single Men.” Journal of Public Economics 95: 1096–1110. https://doi.org/10.1016/j.jpubeco.2011.05.007.
10.1016/j.jpubeco.2011.05.007
Web of Science® Google Scholar
Bargain, O., H. Immervoll, and H. Viitamäki. 2012. “No Claim, no Pain. Measuring the Non-Take-Up of Social Assistance Using Register Data.” Journal of Economic Inequality 10: 375–395. https://doi.org/10.1007/s10888-010-9158-8.
10.1007/s10888-010-9158-8
Web of Science® Google Scholar
Barrios, S., M. Dolls, A. Maftei, et al. 2019. “Dynamic Scoring of Tax Reforms in the European Union.” Journal of Policy Analysis and Management 38: 239–262. https://doi.org/10.1002/pam.22105.
10.1002/pam.22105
Web of Science® Google Scholar
Casas, R. P. 2005. “Minimum Income Standards in Enlarged EU: Guaranteed Minimum Income Schemes.” Transnational Exchange Project. Setting Minimum Social Standards Across Europe, EAPN, Working Paper, 2, 2005.
Google Scholar
Christl, M., S. De Poli, F. Figari, et al. 2024. “Monetary Compensation Schemes During the COVID-19 Pandemic: Implications for Household Incomes, Liquidity Constraints and Consumption Across the EU.” Journal of Economic Inequality 22: 411–431. https://doi.org/10.1007/s10888-023-09596-4.
10.1007/s10888-023-09596-4
Web of Science® Google Scholar
Coady, D., S. Jahan, B. Shang, and R. Matsumoto. 2021. Guaranteed Minimum Income Schemes in Europe: Landscape and Design. International Monetary Fund.
Google Scholar
Collado, D., B. Cantillon, K. Van den Bosch, T. Goedemé, and D. Vandelannoote. 2016. “The End of Cheap Talk About Poverty Reduction: The Cost of Closing the Poverty Gap While Maintaining Work Incentives (ImPRovE Working Paper no. 16/08).” Herman Deleeck Centre for Social Policy, University of Antwerp. https://econpapers.repec.org/paper/hdlimprov/1608.htm.
Google Scholar
Eurofound. 2015. Access to Social Benefits: Reducing Non-Take-Up. Publications Office of the European Union.
Google Scholar
European Commission. 2022a. “Proposal for a Council Recommendation on Adequate Minimum Income Ensuring Active Inclusion.” Brussels, 28.9.2022, COM(2022) 490 Final.
Google Scholar
European Commission. 2022b. “ The 2022 Minimum Income Report.” In Joint Report Prepared by the Social Protection Committee (SPC), vol. 1. Publications Office of the European Union.
Google Scholar
Figari, F., M. Iacovou, A. J. Skew, and H. Sutherland. 2012. “Approximations to the Truth: Comparing Survey and Microsimulation Approaches to Measuring Income for Social Indicators.” Social Indicators Research 105: 387–407.
10.1007/s11205-010-9775-4
Web of Science® Google Scholar
Figari, F., M. Matsaganis, and H. Sutherland. 2013. “Are European Social Safety Nets Tight Enough? Coverage and Adequacy of Minimum Income Schemes in 14 EU Countries.” International Journal of Social Welfare 22: 3–14. https://doi.org/10.1111/j.1468-2397.2012.00885.x.
10.1111/j.1468-2397.2012.00885.x
Web of Science® Google Scholar
Frazer, H., and E. Marlier. 2016. Minimum Income Schemes in Europe. A Study of National Policies 2015. European Commission.
Google Scholar
Frick, J. R., and K. Krell. 2010. “ Measuring Income in Household Panel Surveys for Germany: A Comparison of EU-SILC and SOEP.” SOEPpapers, 265, DIW. https://doi.org/10.2139/ssrn.1540662.
10.2139/ssrn.1540662
Google Scholar
Fuchs, M., K. Gasior, T. Premrov, K. Hollan, and A. Scoppetta. 2020. “Falling Through the Social Safety Net? Analysing Non-Take-Up of Minimum Income Benefit and Monetary Social Assistance in Austria.” Social Policy & Administration 54: 827–843. https://doi.org/10.1111/spol.12581.
10.1111/spol.12581
Web of Science® Google Scholar
Garbuszus, J. M., N. Ott, S. Pehle, and M. Werding. 2021. “Income-Dependent Equivalence Scales: A Fresh Look at German Micro-Data.” Journal of Economic Inequality 19: 855–873. https://doi.org/10.1007/s10888-021-09494-7.
10.1007/s10888-021-09494-7
Web of Science® Google Scholar
Hauser, R. 2008. “ Problems of the German Contribution to EU-SILC – A Research Perspective, Comparing EU-SILC, Microcensus and SOEP.” SOEPpapers, 86, DIW. https://doi.org/10.2139/ssrn.1096524.
10.2139/ssrn.1096524
Google Scholar
Hernández, A., F. Picos, and S. Riscado. 2022. “Moving Towards Fairer Regional Minimum Income Schemes in Spain.” Journal of European Social Policy 32: 452–466. https://doi.org/10.1177/09589287221088174.
10.1177/09589287221088174
Web of Science® Google Scholar
Hernanz, V., F. Malherbet, and M. Pellizzari. 2004. “ Take-Up of Welfare Benefits in OECD Countries: A Review of the Evidence.” In OECD Social, Employment and Migration Working Papers 17. OECD. https://doi.org/10.1787/525815265414.
Google Scholar
Hufkens, T., T. Goedemé, K. Gasior, et al. 2019. “The Hypothetical Household Tool (HHoT) in EUROMOD: A New Instrument for Comparative Research on Tax-Benefit Policies in Europe.” International Journal of Microsimulation 12: 68–85. https://doi.org/10.34196/ijm.00208.
10.34196/ijm.00208
Google Scholar
Immervoll, H., R. Hyee, R. Fernandez, and J. Lee. 2022. “ How Reliable Are Social Safety Nets? Value and Accessibility in Situations of Acute Economic Need.” IZA Discussion Paper, 15232, IZA. https://doi.org/10.2139/ssrn.4114808.
10.2139/ssrn.4114808
Google Scholar
Jara, H. X., and A. Tumino. 2013. “Tax-Benefit Systems, Income Distribution and Work Incentives in the European Union.” International Journal of Microsimulation 6: 27–62.
10.34196/ijm.00076
Google Scholar
Ko, W., and R. A. Moffitt. 2024. “ Take-Up of Social Benefits.” In Handbook of Labor, Human Resources and Population Economics, edited by K. F. Zimmermann, 1–42. Springer International Publishing. https://doi.org/10.1007/978-3-319-57365-6_372-2.
10.1007/978-3-319-57365-6_372-3
Google Scholar
Leventi, C., H. Sutherland, and I. V. Tasseva. 2019. “Improving Poverty Reduction in Europe: What Works Best Where?” Journal of European Social Policy 29: 29–43. https://doi.org/10.1177/0958928718792130.
10.1177/0958928718792130
Web of Science® Google Scholar
Lynn, P., A. Jäckle, S. P. Jenkins, and E. Sala. 2004. “ The Impact of Interviewing Method on Measurement Error in Panel Survey Measures of Benefit Receipt: Evidence From a Validation Study.” In ISER Working Paper Series (No. 2004–28). University of Essex.
Google Scholar
Marchal, S., S. Kuypers, I. Marx, and G. Verbist. 2021. “But What About That Nice House You Own? The Impact of Asset Tests in Minimum Income Schemes in Europe: An Empirical Exploration.” Journal of European Social Policy 31: 44–61. https://doi.org/10.1177/0958928720970134.
10.1177/0958928720970134
Web of Science® Google Scholar
Marchal, S., I. Marx, and N. V. Mechelen. 2014. “The Great Wake-Up Call? Social Citizenship and Minimum Income Provisions in Europe in Times of Crisis.” Journal of Social Policy 43: 247–267. https://doi.org/10.1017/S0047279413000950.
10.1017/S0047279413000950
Web of Science® Google Scholar
Marchal, S., and N. van Mechelen. 2017. “A New Kid in Town? Active Inclusion Elements in European Minimum Income Schemes.” Social Policy & Administration 51: 171–194. https://doi.org/10.1111/spol.12177.
10.1111/spol.12177
Web of Science® Google Scholar
Matsaganis, M., H. Levy, and M. Flevotomou. 2010. “Non-Take Up of Social Benefits in Greece and Spain.” Social Policy & Administration 44: 827–844. https://doi.org/10.1111/j.1467-9515.2010.00746.x.
10.1111/j.1467-9515.2010.00746.x
Web of Science® Google Scholar
Matsaganis, M., A. Paulus, and H. Sutherland. 2008. The Take up of Social Benefits. European Observatory on the Social Situation and Demography: Research Note (2). European Commission.
Google Scholar
Natili, M. 2020. “Worlds of Last-Resort Safety Nets? A Proposed Typology of Minimum Income Schemes in Europe.” Journal of International and Comparative Social Policy 36: 57–75. https://doi.org/10.1080/21699763.2019.1641134.
10.1080/21699763.2019.1641134
Google Scholar
Nelson, K. 2010. “Social Assistance and Minimum Income Benefits in Old and New EU Democracies.” International Journal of Social Welfare 19: 367–378. https://doi.org/10.1111/j.1468-2397.2009.00671.x.
10.1111/j.1468-2397.2009.00671.x
Web of Science® Google Scholar
Nelson, K. 2012. “Counteracting Material Deprivation: The Role of Social Assistance in Europe.” Journal of European Social Policy 22: 148–163. https://doi.org/10.1177/0958928711433658.
10.1177/0958928711433658
Web of Science® Google Scholar
Nelson, K. 2013. “Social Assistance and EU Poverty Thresholds 1990–2008. Are European Welfare Systems Providing Just and Fair Protection Against Low Income?” European Sociological Review 29: 386–401. https://doi.org/10.1093/esr/jcr080.
10.1093/esr/jcr080
Web of Science® Google Scholar
Nelson, K., and R. Nieuwenhuis. 2021. “Towards a New Consolidated Framework for Analysing Benefit Coverage.” Journal of European Social Policy 31: 352–362. https://doi.org/10.1177/0958928721996653.
10.1177/0958928721996653
Web of Science® Google Scholar
Penne, T., T. Hufkens, T. Goedemé, and B. Storms. 2020. “To What Extent Do Welfare States Compensate for the Cost of Children? The Joint Impact of Taxes, Benefits and Public Goods and Services.” Journal of European Social Policy 30: 79–94. https://doi.org/10.1177/0958928719868458.
10.1177/0958928719868458
Web of Science® Google Scholar
Sutherland, H., and F. Figari. 2013. “EUROMOD: The European Union Tax-Benefit Microsimulation Model.” International Journal of Microsimulation 6: 4–26. https://doi.org/10.34196/ijm.00075.
10.34196/ijm.00075
Google Scholar
Tasseva, I. V. 2016. “Evaluating the Performance of Means-Tested Benefits in Bulgaria.” Journal of Comparative Economics 44: 919–935. https://doi.org/10.1016/j.jce.2016.02.003.
10.1016/j.jce.2016.02.003
Web of Science® Google Scholar
Trindade, L. Z., and T. Goedemé. 2020. The Comparability of the EU-SILC Income Variables: Review and Recommendations (Eurostat Statistical Working Papers). Luxembourg: Publications Office of the European Union.
Google Scholar
Van Mechelen, N., and S. Marchal. 2013. “ Struggle for Life: Social Assistance Benefits, 1992–2009.” In Minimum Income Protection in Flux, edited by I. Marx and K. Nelson, 28–53. Palgrave Macmillan UK. https://doi.org/10.1057/9781137291844_2.
10.1057/9781137291844_2
Google Scholar
Van Oorschot, W. 1991. “Non-Take-Up of Social Security Benefits in Europe.” Journal of European Social Policy 1: 15–30. https://doi.org/10.1177/095892879100100103.
10.1177/095892879100100103
Google Scholar
Wang, J., and O. Van Vliet. 2016. “Social Assistance and Minimum Income Benefits: Benefit Levels, Replacement Rates and Policies Across 26 Oecd Countries, 1990–2009.” European Journal of Social Security 18: 333–355. https://doi.org/10.1177/138826271601800401.
10.1177/138826271601800401
Google Scholar

Volume71, Issue1

February 2025

e12728

The Role of Minimum Income Schemes in Poverty Alleviation in the European Union

ABSTRACT

1 Introduction

2 Concept and Main Characteristics of MI Schemes in the European Union

3 The Simulation of MI Schemes in EUROMOD

4 A Methodology to Calibrate the Simulation of MI Schemes

4.1 The Choice of External Data

4.2 The Selection of MI Beneficiaries

4.3 Validation of the “Calibrated Baseline”

5 Assessing the Effectiveness of EU MI Schemes in Poverty Alleviation

5.1 Coverage

5.2 Adequacy

5.3 Poverty-Alleviation Effects

5.4 Evaluating Alternative Calibrations for Selecting MI Beneficiaries

6 Exploring Possible Reforms to Existing MI Schemes

6.1 The Reform Scenarios: Description and Rationale

6.2 What Improvements in the Living Standards of Poor Households?

6.3 How Far From Extreme Poverty Eradication?

7 Conclusion and Discussion

Acknowledgments

Conflicts of Interest

Endnotes

Appendix 1: List of MI Schemes and Validation Results

Appendix 2: Additional Results of the Calibration

Appendix 3: Results of the Effectiveness of Existing Schemes Using the 60% Poverty Line

Appendix 4: Additional Results of the Reform Scenarios

Appendix 5: Sensitivity of Coverage and Adequacy to Different Definitions of MI Protection

References

Figures

References

Information

About Wiley Online Library

Help & Support

Opportunities

Connect with Wiley

The Role of Minimum Income Schemes in Poverty Alleviation in the European Union

ABSTRACT

1 Introduction

2 Concept and Main Characteristics of MI Schemes in the European Union

3 The Simulation of MI Schemes in EUROMOD

4 A Methodology to Calibrate the Simulation of MI Schemes

4.1 The Choice of External Data

4.2 The Selection of MI Beneficiaries

4.3 Validation of the “Calibrated Baseline”

5 Assessing the Effectiveness of EU MI Schemes in Poverty Alleviation

5.1 Coverage

5.2 Adequacy

5.3 Poverty-Alleviation Effects

5.4 Evaluating Alternative Calibrations for Selecting MI Beneficiaries

6 Exploring Possible Reforms to Existing MI Schemes

6.1 The Reform Scenarios: Description and Rationale

6.2 What Improvements in the Living Standards of Poor Households?

6.3 How Far From Extreme Poverty Eradication?

7 Conclusion and Discussion

Acknowledgments

Conflicts of Interest

Endnotes

Appendix 1: List of MI Schemes and Validation Results

Appendix 2: Additional Results of the Calibration

Appendix 3: Results of the Effectiveness of Existing Schemes Using the 60% Poverty Line

Appendix 4: Additional Results of the Reform Scenarios

Appendix 5: Sensitivity of Coverage and Adequacy to Different Definitions of MI Protection

References

Figures

References

Related

Information