Eligibility for the in-work RSA and the Working Poor

Transcription

1 Eligibility for the in-work RSA and the Working Poor Céline Marc and Muriel Pucci Caisse nationale d'allocations familiales (CNAF) First suggested in 2005 by the Commission on Families, Vulnerability and Poverty (Hirsch, 2005), the revenue de solidarité active RSA was introduced across mainland France on June 1 st Designed for people of a working age with a low disposable income, it affects the former recipients of the RMI (social welfare for those not entitled to unemployment benefits) and the API (single parent allowance), as well as being aimed at low-wage workers not receiving any type of statutory minimum welfare benefits. It replaces these minimum benefits, but also the related salary/benefits aggregation scheme 1, and the amount of the in-work RSA gets deducted from the PPE (working tax credit) 2. This new allowance scheme has a three-pronged goal, (i) to remove the threshold effect that can limit the financial benefits of employment for recipients of minimum benefits; (ii) to combat poverty by improving in particular the financial situation of the least well-off workers; and (iii) to simplify the social security system by making it easier to navigate (HCSA, Livre Vert du RSA, 2008). The in-work category of the RSA is presented as a weapon in the fight against in-work poverty. It is way of providing a recurrent income supplement for people who are in employment and should reduce the rate of poverty within the working population. It is therefore directed primarily at family units that are experiencing in-work poverty, as defined by public policy. However, this definition of the working poor, implicitly understood from the RSA eligibility criteria, does not necessarily tie in with the statistical definition of poor workers as adopted by Europe (Eurostat). We will see on the one hand that the RSA uses a broader implicit definition of workers than the European definition and, on the other hand, that the income base criterion used to establish eligibility for the RSA differs from that used to define income poverty, both in terms of the income sources sources considered and the threshold used. This study hopes to compare the notion of the working poor as defined by Eurostat with the population actually eligible for the in-work RSA (i.e. the working poor as defined by the RSA). To do so we will use test cases in terms of family configuration and level of earned income, as well as the Myriade microsimulation model (see box) which can be used to identify both the working poor in the Eurostat sense and those individuals eligible for the in-work RSA (eligible during at least one quarter throughout the year). First, we will show that the definitions of workers and poverty as used by Eurostat and by the RSA change the extent of the problem and the characteristics of the specified population. Secondly, we will demonstrate, using test cases, that the definition of a poor worker as implied by the RSA eligibility criteria does not always coincide with the definition used by Eurostat. Thirdly, using a microsimulation model representative of mainland France, we will quantify these A complete and enriched version of this study will be published in June 2011 in Politiques Sociales et Familiales n More specifically, the three month aggregation period during which welfare benefits can be received on top of earned income still exists. On the other hand, the 1,000 back-to-work bonus paid in the fourth month following return to work and the nine-month aggregation period during which a flat-rate or salary-weighted allowance can be received have been axed. 2 The in-work RSA is meant as an advance on the PPE. Therefore, the annual amount of the in-work RSA (paid monthly) gets deducted from any PPE that may be earned for the following year. 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 1

2 problems which can be explained by the specificity of the RSA eligibility criteria, by fluctuations in income throughout the year, by the specificity of the income sources used to determine RSA eligibility and by the effects of household composition. 1. Working poor as defined by Eurostat and persons eligible for the in-work RSA: two populations that are both quantitatively and qualitatively different In statistical terms, assembling the population of working poor is a complex task since it is based on two major conventions as regards the definition of worker and the definition of poor. The European Union defines worker 3 as an employed individual who has been in the labour force for at least half the reference period 4. This definition therefore excludes individuals who have received earned income but over a shorter period, in particular those who have been mostly unemployed during the year in question. The RSA, however, uses a much broader definition for workers, since an individual only has to have received earned income during the quarter immediately preceding his application for the in-work RSA. Therefore, in terms of this particular allowance, a worker can mean an individual who, in any given year, has received earned income at some point during that year. In addition to those people who are workers in the sense of Eurostat, this definition therefore encompasses every person who has held a very short-term position or who has been mostly unemployed throughout the year. Defining poverty is a regulatory matter influenced by a whole range of conventions specific to each individual society (Verger, 2005). The so-called absolute methods widely used in Anglo- Saxon countries involve consumption-based rules that establish a household s basic needs. European countries prefer to use the notion of relative poverty lines, defined on the basis of a standard of living lower than the majority of the population, and they usually focus on income poverty (as opposed to poor living conditions or subjective poverty 5 ). The criterion used by Eurostat is that of relative income poverty, where poor means an individual whose standard of living is less than 60% of the national equivalised median income. This indicator, seemingly simple in conceptual terms, is nevertheless based on conventional choices when it comes to calculations, what monetary income to consider, what equivalence scale to use to compare standards of living of different sized and structured households and what annual window (calendar year) should be used to define poverty. More specifically, an individual s standard of living is defined by the level of disposable income of the household to which he belongs divided by its equivalent size (number of consumption units). We therefore speak of disposable income per adult equivalent. The equivalence scale used to calculate the number of consumption units in a household (modified OECD scale) is given in table 1. Disposable income is the sum of declared income, financial income, family and housing allowances and statutory minimum benefits, from which get deducted income tax, the CSG 3 See S. Ponthieux (2009) for an in-depth discussion of the main criteria used by institutions to identify the working poor. 4 Due to the problem of incomplete years and figures, the rule used by Eurostat for calculating the European indicator is to have a reference period of at least 7 months (and to pro-rata the employment criterion: 7 months in work out of a 12-month reference period; 6 months out of 11 or 10 etc., down to 4 months in work out of a 7- month reference period). 5 For more information about these types of poverty, see in particular Verger (2005) and Lollivier & Verger (1997). 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 2

3 (Universal Social Security Contribution), the CRDS (Social Debt Repayment Contribution), flatrate taxes and housing tax. One could say that one definition of the poverty line can be implied from the RSA eligibility criteria, namely the income base above which a family is no longer entitled to the RSA (this is of course a simplified definition, bearing in mind the other aims of the RSA outlined above, in particular getting rid of the threshold effect which limits the financial benefits of employment). We will discuss this RSA eligibility threshold later. The income bases that the CAF (Family Allowance Fund) uses to determine a family unit s entitlement to the RSA are those of all people contained in that unit, comprising the worker himself (or herself), any spouse and any dependent child or children under the age of Income base is assessed quarterly and comprises net earned income, key family allowances, housing benefits (limited to a fixed figure) and any other statutory minimum benefits received by the family. The maximum income base up to which the RSA can be received depends on the size of the family, using an implicit equivalence scale described in table 1. Table 1: Modified OECD equivalence scale and equivalence scale implicit in the RSA eligibility criteria Equivalence scale implicit in the RSA eligibility Weighting Modified OECD equivalence scale criteria First adult 1 1 Spouse if over 14: If higher RSA First child * : otherwise: otherwise: 0,3 - if over 14: If higher RSA Second child * : otherwise: otherwise: 0,3 - if over 14: If higher RSA Additional children * : otherwise: otherwise: 0,4 * The higher RSA (single parent) allowance can also be received for one year following separation or adoption when the child is over 3 years of age, but this scenario has not been included in our set of test cases. Therefore, due to differences in the definition of worker and low-income population, the populations of working poor in the sense of Eurostat and of workers eligible for the in-work RSA are in fact quite different. Nevertheless, these two notions share the fact that they combine an individual concept - employment - with a concept defined at household/family level - poverty. They can therefore overlook cases that are highly disadvantaged on the labour market but which are compensated at household or family level (especially women with working spouses or young people still living with their parents); likewise, the population of the in-work poor will comprise a significant proportion of people who have no problems as regards employment (mainly men) but who are the sole breadwinners in their household. However, by being less restrictive regarding the employment criterion, the population of poor workers according to the RSA takes better account of the hardest-hit workers. The data used to assess the extent to which the population of working poor and that of people eligible for the in-work RSA overlap or differ are taken from the Myriade microsimulation model based on the ERFS 7 which is representative of the population in mainland France living in an ordinary household 8. 6 More specifically, a dependent child according to the RSA, if over 16, must have a sufficiently low earned income so as to not reduce the family s RSA allowance, bearing in mind the maximum eligible resource base is raised if there is an additional dependent child. 7 The ERFS (Taxable and Social Revenues Survey) is a statistical comparison of the results from the Continuous Labour Force Survey (data from the 4 th quarter of year N) with the tax files (income declarations) from the GDFiP (Directorate General of Public Finance) for year N and data on contributions received during year N and collected 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 3

4 The Myriade model can be used to assess both household disposable income and the family income base as used by the RSA. Due also to certain gaps in people s employment history over the year, and in order to retain the full population of individuals covered by the ERFS, we have had to make assumptions regarding any missing employment statuses in order to be able to identify workers in the Eurostat sense. This was done based on declared earned income levels over each (calendar) quarter rather than each month in order to limit the possible assumptions for each of the incomplete employment histories. Therefore, the definition of workers in the Eurostat sense is thenceforth taken as individuals who have been employed at least two calendar quarters. It is important to state that there are a lot more families eligible for the in-work RSA than are currently receiving it due to a slow take-up of the scheme. As of September 2010, the number of families actually receiving the in-work RSA represents only 40% of eligible recipients identified by microsimulation. Table 2 shows how the definition of workers as implied by the RSA eligibility criteria applies to more than 2 million more workers than the Eurostat definition. Workers in the RSA sense are more frequently on low earned incomes and are usually poor. The population of poor workers according to Eurostat therefore accounts for 6% of workers in the European sense, a total of 1.4 million individuals. Since the proportion of poor individuals among those who have received earned income during the year is in fact higher, we can count nearly 1.9 million poor people classed as workers for the purposes of the RSA. Table 2: Comparing the two concepts of employment and poverty Workers according to the EU (at least two quarters in employment) Workers according to the in-work RSA (positive earned income) 27,2 Population size (in thousands) Percentage of population of working age 25,1 78% 85% Percentage of earned incomes beneath the poverty line 16% 22% Percentage of poor 6% 7% Percentage eligible for in-work RSA* 8% 9% Source: Myriade, 2010 scales, ERFS 2007, authors calculations Scope: Individuals aged not including students or retired persons. Head of household not a student and non-negative declared income. * Individuals whose family income base is within the in-work RSA eligibility bracket for at least one calendar quarter during the year. We see that the percentage of individuals whose family income base fell within the in-work RSA eligibility bracket for at least one calendar quarter was 9% among the population of individuals who have worked at some point during the year (nearly 2.5 million) compared to only 8% from among the population of workers in the Eurostat sense (2.1 million). Table 3 gives some figures relating to the working poor in the Eurostat sense compared to individuals considered workers by the RSA and entitled to the in-work RSA, whether head of the via the CNAF (National Family Allowance Fund), the CNAV (National Retirement Pension Fund) or the CCMSA (Central Fund for the Agricultural Mutual Insurance Scheme). 8 In keeping with other poverty studies, we have limited our analysis to households where the head of the house is not a student and whose declared income is positive or zero. 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 4

5 household, spouse of the head of household or dependent child. These populations need to be compared with the population of workers in the Eurostat sense and that of all people of working age (not including students and retired persons) who have received earned income at some point during the year. Table 3: Working poor figures calculated using different criteria Workers according to Eurostat (at least two quarters in employment) Workers according to the In-work RSA (positive earned income) Working Poor Total Eligible for Inwork RSA* Total Population size (in thousands) 1,4 25,1 2,5 27,2 Sex Proportion of women 42% 47% 49% 48% Situation as regards the labour market Intermittent worker 34% 12% 29% 14% Part time 29% 15% 29% 16% Year-round employment 70% 88% 68% 82% At least two quarters without employment 12% 4% 11% 8% Family configuration Single no children 22 % 23 % 22 % 23 % Single with one child 6 % 3 % 11 % 3 % Single with two children 4 % 1 % 4 % 2 % Single with three or more children 2 % 0 % 1 % 1 % Couple no children 14 % 26 % 16 % 26 % Couple with one child 15 % 19 % 21 % 18 % Couple with two children 20 % 20 % 17 % 19 % Couple with three or more children 17 % 8 % 9 % 8 % Source: Myriade, 2010 scales, ERFS 2007, authors calculations Scope: Individuals aged not including students or retired persons, who have been employed at some point during the year. * Eligible for In-work RSA means individuals who have received earned income at some point during the year and whose income base is sufficiently low, with regard to the RSA criteria, to be entitled to the In-work RSA for at least one quarter during that year. By comparing workers in the Eurostat sense and individuals who have worked at some point during the year (columns 2 and 4), we see that these two populations are fairly similar in terms of make-up. However, workers according to Eurostat are less likely to be women (possibly due to the fact that the European criteria require relative job stability that women are less likely to enjoy), experience greater job stability and are less likely to be single parents. Within the population of workers according to Eurostat (columns 1 and 2), the in-work poor are less likely to be women (42% vs. 47%). They experience less job security than the average, are more likely to be in part-time or non-full-time employment, are less often in a couple and usually have two or more dependent children. Among those individuals who have worked at some point during the year (columns 3 and 4), women account for 49% of eligible for the In-work RSA and 48% of overall. Those eligible are in more precarious situations than the average and are more likely to be part-time workers or in non-full-time employment. Single parents with one dependent 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 5

6 child are highly over-represented among the eligible population (11% vs. 3%) whereas childless couples are highly under-represented (16% vs. 26%). If we look just at low-income workers according to the relative income poverty criteria and according to the RSA eligibility criteria (columns 1 and 3), there are almost twice as many poor workers in the European sense than individuals who have worked and who are eligible for the Inwork RSA (nearly 2.5 million vs. 1.4 million). If we consider RSA eligibility to be an implicit definition of one form of poverty (without forgetting that this is a simplified view, given the other aims of the RSA scheme, especially the hope of removing the threshold effect), this would appear to show that the In-work RSA is based on a broader notion of in-work poverty. In fact, those persons who have received some earned income during the year and who are eligible for the inwork RSA are usually more secure than poor workers in the Eurostat sense, even though they are less so than the average worker. Those eligible for the in-work RSA are more likely to be women and individuals with at most one child (single parent or couples) than the working poor according to Eurostat. These initial qualitative results expose the extent to which Eurostat s population of poor workers and the group of people who have received some earned income and are therefore eligible for the in-work RSA during the year do not in fact overlap. Now let us clarify the differences between these two populations by considering two questions: To what extent are those eligible for the RSA also poor workers according to Eurostat? To what extent are poor workers according to Eurostat also eligible for the RSA? In order to answer these questions, we will first of all, using test cases in terms of family configuration, compare the level of earned income needed to get out of poverty and that needed to exceed the RSA eligibility threshold. We will see that the salary level at which an individual can and cannot claim the in-work RSA and at which he is or is not poor depends on his family configuration. The Myriade model will then allow us to quantify and clarify this diagnosis. 2. Test case analysis of those eligible for the in-work RSA and of poor workers in the statistical sense In order to identify which workers are and are not eligible for the in-work RSA and to determine to what extent these two categories of workers are and are not poor according to Eurostat, we have created test cases (2010 legislation) based on family configuration - single; couples where one or both partners works (earned income being shared equally between the spouses); no children or with 1, 2 or 3 children - and the level of earned income the family receives. To be able to compare families with different numbers of workers and different compositions, earned income is expressed as consumption units (CU). After comparing the equivalence scales used to calculate standard of living and for RSA eligibility (table 1), we decided to distinguish between families with and without a child under 3 or over 14 years old. Children under 3 can create entitlement to the basic allowance under the PAJE (Early Childhood Benefits) Programme, and children over 14 to an increase in family benefits (not included by the RSA in a family s income base). Graph 1 shows us: The RSA exit points, in other words the earned income per CU above which the family is no longer eligible for the in-work RSA. 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 6

7 The poverty exit points, in other words the minimum earned income per CU that a family needs for its standard of living to be at least equal to the poverty line, account taken of any social allowances and statutory minimum benefits received. The poverty exit points prior to the in-work RSA, in other words the minimum earned income per CU that a family needed for its standard of living to be at least equal to the poverty line before the in-work RSA was introduced. We remember that before the RSA was introduced, families benefitted from the PPE (working tax credit). With the new system, the in-work RSA is meant as an advance on the PPE, and the annual amount of the in-work RSA gets deducted from any PPE that may be earned for the following year; this can either cancel out the PPE completely, or mean the RSA recipient gets a top-up via the PPE scheme. Test cases suppose that a person s situation does not change over the year and calculate the amount of PPE top-up (positive or zero) for families receiving the inwork RSA. For some family configuration/earned income combinations where a positive PPE topup applies, the standard of living before and after the RSA is the same. Graph 1: Eligibility for in-work RSA and poverty, broken down by family configuration Families with children aged between 3 an 14 years Families with one child aged under 3 years, and any other being between 3 an 14 years Couple 2 working 3 children Couple 2 working 2 children Couple 2 working 1 child Couple 2 working 0 children Couple 2 working 3 children Couple 2 working 2 children Couple 2 working 1 child Couple 2 working 0 children Couple 1 working 3 children Couple 1 working 2 children Couple 1 working 1 child Couple 1 working 0 children Couple 1 working 3 children Couple 1 working 2 children Couple 1 working 1 child Couple 1 working 0 children Single 3 children Single 2 children Single 1 child Single 0 children 0% 50% 100% 150% Earned income per CU (as % of minimum wage) Single 3 children Single 2 children Single 1 child Single 0 children 0% 50% 100% 150% Earned income per CU (as % of minimum wage) Families with one child aged over 14 years, and any other being between 3 an 14 years Couple 2 working 3 children Couple 2 working 2 children Couple 2 working 1 child Couple 2 working 0 children RSA exit point Poverty exit point before introduction of in-work RSA Couple 1 working 3 children Couple 1 working 2 children Couple 1 working 1 child Couple 1 working 0 children Poverty exit point Poverty line (90% of minimum wage) Single 3 children Single 2 children Single 1 child Single 0 children 0% 50% 100% 150% NB : Earned income of couples where both are working is shared equally between the spouses Earned income per CU (as % of minimum wage)% du Smic) 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 7

8 Comparing the poverty exit points before and after introduction of the in-work RSA shows us, for each type of family configuration, the salary level at which the families have actually been able to get out of poverty thanks to the in-work RSA. Comparing poverty exit points (before or after introduction of the in-work RSA) and RSA exit points, we can identify the salary levels at which families are entitled to the in-work RSA even though their standard of living is in fact above the poverty line. More conclusions can be drawn from these test cases. First, the RSA exit point falls with number of children, except for the first child of couples whose children are all aged between 3 and 14. For example, in the category of Couple with one child under the age of 3 and any other children between 3 and 14 years, the RSA exit point is approximately 115% of the minimum wage per CU (approx 1,800 for 1.5 CU) if they have no children; 70% of the minimum wage per CU (approx 1,330 for 1.8 CU) if they have one child under 3; approximately 60% of the minimum wage per CU ( 1,370 for 2.1 CU) if they have two children one of which is under 3; and 40% of the minimum wage per CU ( 1,010 for 2.4 CU) if they have three children one of which is under 3. This fall in the RSA exit point can be explained by the fact that the RSA, as a weighted allowance, accounts for differences in allowance entitlements between families. Therefore, the amount of RSA that a family can claim will fall if its family (and housing) allowance entitlement increases. Secondly, and for a similar reason, the salary per CU at which a family can get out of poverty thanks to the existence of the in-work RSA falls when the number of children increases, and is even non-existent for families with three children, regardless of their age. This can be explained at once by the relatively high value of family benefits, the CF (Large Family Supplement) and by the existence, for these families, of a PPE supplement granted at the same level as the RSA exit point. Thirdly, for most 3-children families and for couples where one spouse is working and they have two children, one of which is under 3, the RSA exit point is less than the poverty exit point, implying that for certain salary levels, these families can be poor in the sense of income poverty but still not be eligible for the in-work RSA. Among the 3-child families, only couples where both are working and who have only children between 3 and 14, or individuals with three children one of which is under 3 have an RSA exit point higher than the poverty exit point. Finally, the fourth conclusion to be drawn from these test cases is that seemingly for some family configurations, the RSA exit point is above the poverty line. In other words the RSA is open to families who would not be poor in the statistical sense even though they had no resources other than their earned income (neither housing benefits, nor family allowances). This applies to childless families, couples with one child (except if aged under 3) and couples with two children aged between 3 and 14 years. In summary, these test cases clearly show that the working poor in the statistical sense and those eligible for the in-work RSA (working poor in the sense of public policy) are not one and the same population. Some poor workers in the statistical sense are not entitled to the RSA 9 whilst some of those who are eligible are not, statistically speaking, poor, or have even gotten out of poverty thanks to the in-work RSA. The main disadvantage of these test cases is that they do not take account of the representativeness of the scenarios used. By generalising the transfer scale deduction to a representative sample of ordinary households from mainland France, the Myriade microsimulation model can quantify the various scenarios and give reasons for the partial non- 9 We remember that individuals under 25 with no dependent children are not eligible for the in-work RSA. They may, however, from September 2010 and subject to certain conditions regarding previous employment, be entitled to the Young Persons RSA, which is not covered by this report. 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 8

9 overlap between Eurostat s working poor and those individuals who have worked at some point and therefore eligible for the in-work RSA during the year. 3. Using Myriade to analyse differences between Eurostat s working poor and individuals who have worked and are eligible for the in-work RSA during the year Table 3 shows that the population of people who have received earned income and who are eligible for the in-work RSA is almost 2 times larger than that of the working poor as defined by Eurostat, which clearly means that some eligible persons are not in fact poor workers according to the European definition. At the same time, our analysis of test cases revealed that for some family configurations and for some salary levels, certain poor workers are being excluded from entitlement to the income supplement. In this section we will try to quantify these issues and clarify the reasons why a proportion of the people eligible for the in-work RSA are not in fact poor workers (section 3.1), and on the other hand why some poor workers are unable to claim the supplement (section 3.2) Which family units are eligible for the RSA but are not in fact families of working poor as defined by Eurostat? Our test case analysis has already allowed us to identify certain salary levels at which non-poor workers, according to the European criterion, are nevertheless eligible for the in-work RSA. In fact, for most family configurations, the RSA exit point is above where the poverty exit point was before the RSA was introduced. However, these test cases suppose the worker s employment situation remains the same throughout the year and that the family s income base comprises only their salaries, housing benefits, family allowances, the Large Family Supplement, Early Childhood Benefits, the ASF (Family Support Allowance) and the RSA 10. In addition, the test cases reflect families that have received earned income, but not individual workers, according to either the European definition or the wider definition used for RSA eligibility. Thanks to the microsimulation using the Myriade model, it has been possible to analyse the situations of workers taking into account both possible fluctuations in their income throughout the year and their overall income. Table 4 gives an analysis of families eligible for the in-work RSA at some point during the year, distinguishing between those who comprise at least one worker according to Eurostat. First of all, it seems that 89% of all family units eligible for the in-work RSA comprise at least one worker in the European sense. Also, 36% of all eligible family units are families of working poor and 9% contain at least one worker in the European sense who has gotten out of poverty thanks to the in-work RSA. This latter subset of workers can be described as in-work RSA-dependent'. To these poor and dependent family units, we can add the 9% of eligible family units which have received earned income and which are poor but in which no one member has had sufficient employment to be classed as a worker by Eurostat. 10 The income base on which the RSA is calculated also includes any disabled persons allowance, elderly worker income supplement, pension benefits or maintenance payments received, as well as the PAJE elective employment supplement, none of which are taken into account by the model cases. 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 9

10 Table 4: Family units eligible for the in-work RSA Family units containing at least one worker according to Eurostat Family units containing no workers according to Eurostat All eligible family units Poor 36% 9% 45% in-work RSA-dependent 9% 0% 9% Neither poor nor dependent 44% 2% 46% Total 89% 11% 100% Population size (in thousands) 1, ,917 Source: Myriade, 2010 scales, ERFS 2007, authors calculations Scope: Families in the sense of the RSA who, based on the eligibility criteria, are entitled to receive the in-work RSA, including those benefitting from the 3-month salary/benefits aggregation period. This therefore leaves 46% of eligible family units who are neither poor nor RSA-dependent (approx 0.9 million families), of which 44% comprise at least one worker in the European sense (approx 0.84 million families). Before looking at how RSA-eligible family units break down by family configuration and before assessing how some families can be eligible without being either poor or dependent, it will help to determine how the in-work RSA would affect poverty if all eligible units claimed it. Table 5: impact of quarterly in-work RSA on poverty pre in-work RSA post in-work RSA Poverty rate* 12.6% 11.9% Poverty intensity 16.1% 14.8% Individuals getting out of poverty 414,500 RSA family units getting out of poverty 192,800 Source: Myriade, 2010 scales, ERFS 2007, authors calculations Scope: Households where the head of the household is not a student and whose declared income is positive or zero. * The poverty rate is calculated as 917/month, based on living standards prior to the in-work RSA. Table 5 shows that the in-work RSA, if it were fully subscribed, would reduce the poverty rate by 0.7 percentage points and poverty intensity 11 by 1.3 points. It would allow 414,500 individuals (belonging to 192,800 RSA family units) to get out of poverty. We note that among the RSA family units that could get out of poverty thanks to the in-work RSA, 170,700 are directly eligible for the in-work RSA and the rest would just benefit from it since they belong to a household that contains an eligible family unit. 11 The poverty rate is the percentage of individuals living in households (with non-negative declared incomes and whose head of household is not a student) whose standard of living is below the poverty line, equal by definition to 60% of the median standard of living. Poverty intensity measures the median difference between the poverty line and the standard of living of poor individuals. 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 10

11 Graph 2 gives a breakdown of RSA-eligible family units by family configuration. We see that more than a quarter of family units eligible for the in-work RSA are single persons. They are followed by couples with one child (17% of eligible families), then single parents with one child and couples with no or two children (just under 15%). This partly echoes the breakdown of workers by family configuration. Graph 3 gives a more detailed view of how the in-work RSA is divided between the various types of family configurations (the only other differentiating factor being the distribution of earned income specific to each configuration). The graph shows, for each family configuration, the proportion of workers in the Eurostat sense who are eligible for the RSA from among all workers in the Eurostat sense. It tells us that workers with children are more often entitled than workers with no children. In particular, it seems that more than one quarter of single parent workers with one dependent child are eligible, as are approx 20% of single parent workers with two or more dependent children. The second thing we see is that single workers are more likely to be eligible than those in couples, even when they have the same number of dependent children. Thirdly, it appears that the likelihood of being entitled to the in-work RSA whilst remaining statistically poor increases with number of dependent children, regardless of whether the worker is in a couple or otherwise. The fourth finding, and contrary to what the test cases suggested, is that the in-work RSA can allow workers from all the types of family configuration to get out of poverty. The actual probability of being able to get out of poverty thanks to the in-work RSA does however differ depending on family situation: whereas approx 4% of single workers with one dependent child get out of poverty thanks to the in-work RSA, the figure is only 2% for single workers with two children, 1% for childless single workers, less than 0,5% for childless couples and only 0.2% for single workers with three or more children. Finally, the fifth conclusion is that all family configurations create the possibility of being eligible for the in-work RSA even with a standard of living above the poverty line. 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 11

12 Graph 2: Composition of family units eligible for the in-work RSA Couple 3 children Couple 2 children Couple 1 child Couple no children Single 3 children Single 2 children Single 1 child Single no children 0% 5% 10% 15% 20% 25% 30% As % of all eligible family units Poor Dependent Neither poor nor dependent Source: Scope: NB: Myriade, 2010 scales, ERFS 2007, authors calculations Family units eligible for the in-work RSA The categories single 3 children and couple 3 children also include families with three or more dependent children. Graph 3: Workers, as defined by Eurostat, eligible for the in-work RSA, broken down by family configuration Couple 3 children Couple 2 children Couple 1 child Couple no children Single 3 children Single 2 children Single 1 child Single no children 0% 5% 10% 15% 20% 25% 30% As % of workers (as defined by Eurostat) with the same family configuration Poor Dependent Neither poor nor dependent Source: Myriade, 2010 scales, ERFS 2007, authors calculations Scope: Workers as defined by Eurostat NB: The categories single 3 children and couple 3 children also include families with three or more dependent children. 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 12

13 We now have to look at the reasons why 0.84 million families of workers (in the European sense) who are eligible for the in-work RSA are in fact neither poor nor dependent (44% of eligible families). In order to quantify the role played by the RSA eligibility criteria (adjusted using the test cases) in creating this problem, we have assessed the influence of certain other reasons, which are nonmutually-exclusive, why some families of RSA-eligible workers can be neither poor nor RSAdependent (graph 4). A family-based approach that can overlook certain members of the household who are not included in the RSA family unit: it is possible for some members of the larger household who are not counted as part of the RSA family unit to have high incomes, such that the overall household s standard of living is above the poverty line, whereas the family unit is still considered poor by the RSA. 29% of eligible families of workers who are neither poor nor dependent (0.24 million families) belong to households comprising several family units, as defined by the RSA. For three quarters of these families, the family income base that is taken into account when assessing RSA eligibility plus the amount of the Basic RSA received is under the poverty line: they are therefore classed as poor from the point of view of an RSA family unit. The full amount of the families disposable income is not taken into account: as we have already pointed out, the income base on which RSA eligibility is determined includes only some of the possible family benefits; the AEEH (Disabled Child Education Allowance), the ARS (School Child Allowance), the PAJE birth grant and part of the PAJE basic allowance during the year of the child s birth are excluded. In addition, taxes are not taken into account. If we only count households that comprise a single family unit in the sense of the RSA, we have approx 120,000 families (14% of the total) whose family income base used to calculate RSA eligibility plus the Basic RSA is less than the poverty line. Different reference periods used to identify low-income families: poverty is assessed based on a family s average standard of living over a year, whereas RSA entitlement is calculated on a family s income base over just one quarter. Therefore some families are only eligible to receive the in-work RSA for part of the year and can have, for the rest of the time, an income base that is too high for entitlement meaning that, on average over the year, their standard of living is higher than the poverty line. If we look just at households comprising one single family unit in the sense of the RSA, we count approximately 0.21 million families of workers which are neither poor nor dependent who are eligible for the in-work RSA for part of the year even though the family income base used to calculate RSA eligibility plus the basic RSA is higher than the poverty line (25% of the total). Finally, the effects of the eligibility criteria: in total, 64% of working families that are neither poor nor dependent and who are eligible for the RSA are not considered poor if we look just at the family income base used to calculate the RSA. If we exclude families that belong to complex (compound) households and those who do not receive the in-work RSA throughout the whole year, there are still approximately 0.27 million family units (32% of the total) of neither poor nor dependent workers whose RSA eligibility can be explained entirely by the nature of the eligibility criteria. 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 13

14 Graph 4: Reasons for eligibility of working families that are neither poor nor dependent 840,000 families of workers (as defined by Eurostat) eligible fo the in-work RSA but who are neither poor nor dependent 29 % belong to complex (compound) households (0.24 million) 71 % belong to simple (single) households (0.6 million) 22 % on low RSA resources * 7 % on high RSA resources 14 % on low RSA resources 57 % on high rsa resources 32 % eligible the whole year criteria effect alone 25 % eligible for some of the year Source: Myriade, 2010 scales, ERFS 2007, authors calculations Scope: Family units that contain at least one worker as defined by Eurostat and which are eligible for the in-work RSA, but whose standard of living without the in-work RSA is higher than the poverty line. * Low RSA resource means families that would be considered poor if disposable income meant only the annual family income base (and not that of the household) used to calculate RSA eligibility plus the Basic RSA Who are the poor workers (in the Eurostat sense) who are not eligible for the in-work RSA? Table 6 shows that 37% of workers who would be considered poor according to European criteria before adding the in-work RSA (poor or RSA-dependent workers) are not entitled to the in-work RSA, amongst which 8% are under-25s poor workers. Table 6: Eligibility for in-work RSA of poor workers in the Eurostat sense before adding the in-work RSA Population size (in % thousands) Eligible for in-work RSA 1,002 64% Under 25 non eligible 122 8% Other non eligible % Population size (in thousands) 1, % Source: Myriade, 2010 scales, ERFS 2007, authors calculations Scope: Poor workers in the Eurostat sense when disposable income is calculated before adding the in-work RSA 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 14

15 Graph 5 shows the percentages of poor workers in the European sense before the in-work RSA, depending on whether they are eligible for this allowance, broken down by family configuration. Graph 5: Poor workers before the in-work RSA, both eligible and not eligible for the inwork RSA Couple 3 children Couple 2 children Couple 1 child Couple no children Single 3 children Single 2 children Single 1 child Single no children 0% 5% 10% 15% 20% 25% As % of all poor workers before in-work RSA according to the European definition Other eligible Young eligible Other non eligible Young non eligible Source: Myriade, 2010 scales, ERFS 2007, authors calculations Scope: Workers who are poor in the Eurostat sense, before the in-work RSA, who may be eligible as head of family, spouse of head of family or dependent child. NB: The categories single 3 children and couple 3 children also include families with three or more dependent children. We find that, for all family configurations, there are poor workers not entitled to the RSA. Whereas the test cases suggested that all poor workers without children were eligible for the in-work RSA, our microsimulation analysis reveals that a significant proportion of these workers are not. This is in particular explained by the non-eligibility of the under-25s (and their spouses) without dependent children (this report does not take into account the extension of the RSA scheme, in September 2010, to include the under-25s, subject to previous employment). This is not however the only reason. Whether we look at single childless workers or workers living in a couple, just over half of those not eligible are over 25. If take just the families with dependent children, we see that the proportion of those not eligible increases with number of children. This proportion goes from approximately one fifth for families with one single child, to more than one third for those with three or more children. To try to better understand the reasons why 573,000 poor workers in the Eurostat sense before the in-work RSA could be excluded from this allowance, and to quantity how much of this is caused by the eligibility criteria, we have assessed the influence of other possible factors. The first reason, which we have already mentioned, is related to the de facto ineligibility of the childless under 25s. Nearly 122,000 poor workers not eligible for the in-work RSA are young people excluded from the scheme (see Graph 5). 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 15

16 Let us now consider the 451,000 other poor workers not entitled to the in-work RSA, i.e. those aged over 25 or with dependents. Beyond the effects of the eligibility criteria which, as we saw with the test cases, can exclude certain poor workers from large families, the potential reasons for non-eligibility relate at once to what unit is used (family unit vs. household), the specificity of the income base used in each case and on the income reference period used. Graph 6: Reasons for ineligibility of poor workers before in-work RSA 573,000 poor workers (according to the Eurostat) before in-work RSA not eligible for the in-work RSA 451,000 individuals aged over 25 years or with dependent children 122,000 individuals aged under 25 years with no dependent children 10 % belong to complex households (approx 46,000) 90% belong to simple households (approx 405,000) 5 % on low RSA resources * 5 % on high Rsa resources 67 % on low Rsa resources 23 % on high RSA resources 39 % receive earned ncome throughout the year criteria effect alone 28 % alternate betwen quarters with and without earned income Source: Myriade, 2010 scales, ERFS 2007, authors calculations Scope: Poor workers according to Eurostat before the in-work RSA. * Low RSA resources means families that would be considered poor if disposable income meant only the annual family income base (and not that of the household) used to calculate RSA eligibility plus the Basic RSA. A family-based approach that can overlook certain members of the larger household who do not belong to the RSA family unit: a family unit can have a relatively large resource base but be required to share it with other family units in the same household whose income is low or perhaps even non-existent. This is in particular the case of households that contain one family unit in the sense of the RSA and one or more adult children who remain objectively dependent on their parents even if they are not so in the eyes of the RSA. Therefore, nearly 10% of poor workers not eligible for the in-work RSA (approximately 46,000) belong to households comprising several family units as defined by the RSA. For one half of these families, their RSA income base plus the basic RSA is higher than the poverty line, and they are not therefore poor from the point of view of an RSA family unit. 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 16

17 Not all a family s disposable income is being taken into account: as we have already discussed, the income base for the RSA, unlike disposable income, does not take into account income tax or housing tax which can bring down a family s standard of living. If we look only at households comprising one single family unit in the sense of the RSA, we count just over 104,000 non-eligible poor workers (approx 23% of the total) who are not poor from the point of view of their RSA income base (income base taken into account for determining RSA eligibility plus the basic RSA is higher than the poverty line). A different reference period is used to identify low-income families: fluctuations in income throughout the year can explain why a worker is considered both non-eligible and non-poor for the quarters in which he receives earned income, but poor (including in terms of his RSA resource base) for the year as a whole. Looking just at simple households, approximately 124,000 poor workers not eligible for the in-work RSA (28% of the total) only have an earned income for part of the year and their income base used to determine RSA eligibility is below the poverty line. Finally, the eligibility criteria effect: in total, 72% of poor workers not entitled to the RSA (325,000 people) are also poor when we consider just the family income base used to determine such entitlement. If we exclude poor workers who belong to compound households and those who experience strong fluctuations in earned income, there are still about 177,000 poor workers (39%) whose non-eligibility is explained entirely by the eligibility criteria. Conclusion This study shows that the eligibility criteria for the in-work RSA only partially cover all the families of poor workers in the statistical sense used by Eurostat (something that can be explained in particular by the other aims being pursued). On the one hand, more than a third (36%) of family units eligible for the in-work RSA are poor working families and a further 9% would be so were it not for the in-work RSA (RSA-dependent workers), meaning only approximately 45% of family units are considered poor working families both from a statistical point of view and in terms of public policy definition. On the other hand, only 64% of the working poor in the European sense before the in-work RSA are eligible for this allowance. The differences in the two definitions of working poor can be explained both by the RSA eligibility criteria and by differences between the two reference periods used (a year for poverty and a quarter for the RSA), what living unit is used (household for poverty and family unit for the RSA), and what the income base includes (disposable income encompasses income as well as additional deductions). These two definitions of the working poor do not therefore coincide but do allow us to isolate populations with a range of vulnerabilities that can accumulate: low annual salary, part-time employment, alternation between periods with and without work as well as other family factors (dependents, spouse s income base). 31 March 1 April 2011 Peer Review Building the tools to fight in-work poverty, France 17