by Hélène Périvier OFCE-PRESAGE

The recent study published by France Strategy on the sharing of the costs of children after a separation has caused a stir (see in particular Dare feminism, Abandoning the family, as well as SOS Papa [all in French]). The study analyses the changes in the standard of living of both the former spouses, taking into account the interaction between the indicative scale for child support and the tax-benefit system. This approach is stimulating, as it endeavours to see whether the redistribution effected through the welfare state fairly and equitably deals with the costs of the child borne by each former spouse.

It is reported that after separating, the living standards of the two former partners fell sharply. In addition, simulations of typical cases “indicate that as a result of applying the scale [the indicative reference scale provided to judges] under existing social and tax legislation, the care of children causes a significantly greater sacrifice in the standard of living of the non-custodial parent than of the custodial parent”. In other words, separated fathers are making a greater sacrifice in their standard of living than are the mothers, if the judge were to apply the indicative scale to the letter. But according to the Ministry of Justice the scale is not applied by judges, as both situations are always very specific. So the study looks at what the standard of living of the separated parents would be if the scale were applied, and not at their actual standard of living. However the table of results presented in the note on the front page is titled, “Estimating the loss of living standards incurred by the parents of two children (as a percentage compared to the situation with no child, calculation net of state aid)”. Someone reading this quickly could easily think this was the real situation of separated parents.

Even though the study is based on the scale for support payments and not on the decisions of the judges themselves, it raises a relevant question. But the results are weakened by significant methodological problems: the concept of the sacrifice in the standard of living does not take into account the gender division of labour and its impact on mothers’ careers; the typical cases highlighted are not necessarily representative (in particular concerning marital status prior to separation); using the equivalence scales [1] leads to conflating the “household standard of living” and “the individual standard of living”; and finally, an approach based on maintaining the child’s standard of living would have led to a completely different result. Ultimately, proposing the micro-simulation model as an aid to the judges’ decision-making seems somewhat premature in light of these criticisms.

On the concept of “a sacrifice in the standard of living” 

In all the cases simulated, the separated parents’ living standards go down relative to their situation as a couple (assuming unchanged income). This result is consistent with other recent work, such as Martin and Périvier, 2015; Bonnet, Garbinti, Solaz, 2015; and the report of France’s Family Council (the HCF). A separation is costly for both parents due to the loss of economies of scale (e.g. two homes are needed instead of one, etc.). In addition to the decline in living standards for each parent, the authors calculate the “sacrifice in living standards” experienced by the parents after the separation.

The “living standard sacrifice” is supposed to be calculated by comparing the cost of the child to the disposable income that the parent would have had if there were no child. However, the living standard sacrifice made by the mother with custody of the child (or the father, respectively) is actually calculated by comparing the child’s cost with the standard of living of a single woman without children with the same salary level as the separated mother (and the same for the father).

This method cannot be used to estimate the “living standard sacrifice”, since forming a couple and a family are accompanied by a gender division of labour, which has been widely documented in the literature and which implies that the separated wife has a salary level, and more generally a career, that is different from what she would have had if she had remained single with no children. If a woman senior executive living in a couple stops working in order to look after the children and then the couple separates, the concept of the “living standard sacrifice” would imply a significant gain in the quality of life for this woman, since the cost of the children would be relative to the RSA minimum income, whereas she would have received a higher salary if she had not had children because she would have continued to work.

In other words, the proper counterfactual, that is to say the situation with which we must compare the level of the separated parent so as to assess the living standard sacrifice that she (or he) suffers, should be the income that the woman (or man) would have had when separated (taking into account their individual characteristics) if she (or he) had not entered a couple and if she (or he) had not had children. By doing this, the calculations would have led to a significantly greater sacrifice by the woman than that calculated in the study. Here we see the need for an economic approach that integrates the behaviour of agents, compared with an accounting approach.

Atypical typical cases?

The authors used the micro-simulation model Openfisca to simulate different situations and assess the loss in living standard by each former spouse after the separation.

The typical cases are used to understand the complex interactions between the tax-benefit system and, for the subject matter here, the indicative scale of child support payments. The criticism usually made of typical case studies is that they do not reflect the representativeness of the situations simulated: so to avoid focusing on marginal cases, data is added about the frequency of the situations selected as “typical”. With respect to the distribution of income, in three-quarters of the cases the women earn less than their male partners (Insee). What would be needed is to look at the distribution of income between spouses before the break and see what are the most common cases and then to refine the operation by retaining only those cases where the judge sets a support payment, i.e. in only 2 out of 3 cases (Belmokhtar, 2014).

Likewise, focusing on the case of a couple with two dependent children is not without consequences[2], since with only one dependent child the amount of family benefits falls, meaning that the social benefits received by the mother would be lower (in particular the family allowance is paid only starting from the second child) as would her standard of living. Statistics provided by the Ministry of Justice indicate that the average number of children is 1.7 in the case of divorces and 1.4 in the case of common-law unions (Belmokhtar, 2014).

Finally, nothing is said explicitly about the marital situation prior to the separation: marriage or common-law?

– Either the authors are considering married couples. In this case, if the salaries of the ex-spouses are different (case 4 described as “Asymmetry of income”), how is the loss of France’s marital quotient benefit (quotient conjugal) distributed? After divorce, the tax gain resulting from joint taxation is lost: the man then pays a tax amount based on his own salary and no longer on the couple’s average salary. This additional tax burden hits his living standard, and the “living standard sacrifice” calculated for the divorced father would then partly reflect the loss of this marital quotient benefit, and not the cost arising from the expense of a separated child.

– Or the authors consider only common-law couples, which seems to be the case given the vocabulary used – “separation, union, separated parents, etc.” – but then this brings back the criticism about the representativeness of the typical cases, since more than half of the court decisions regarding the children’s residence are related to divorces (Carrasco and Dufour, 2015). Moreover, the support payments set by the judge are all the more distant from the scale in the case of a separation and not a divorce, which limits the scope of the study.

On the proper use of equivalence scales

Equivalence scales are used to compare the living standards of households of different sizes, by applying consumption units (CU) to establish an “adult equivalent”. These scales are based on strong assumptions that do not allow the use of this tool in just any old way, i.e.:

– that individuals belonging to a single household pool their resources in entirety;

– that people belonging to the same household have the same standard of living (the average standard of living is calculated by dividing the total household income by the number of household CUs). This assumption flows from the first; the standard of living is equated with well-being.

Equivalence scales give an estimate of the additional cost linked to the presence of an additional person in a household. They say nothing about the way in which resources are actually allocated within the household. This is due to the hypothesis that resources are pooled, which is questionable (see in particular Ponthieux, 2012) and which leads to attributing the household’s average standard of living to each individual member. A couple has 1.5 CU. In fact, a couple A in which the man earns 3 times the minimum wage (SMIC) and the woman 0 times the SMIC would have the same standard of living as a couple B in which both earn 1.5 times the SMIC. This method can be used to compare the average living standards of two households, but not the living standards of the individuals who compose them. The woman in couple B probably has an individual standard of living that is higher than the woman in couple A, due to her greater bargaining power given the equal wages earned. So comparing the average living standards of the couple with the living standards of the individuals when the couple separates is misleading.

Likewise, to assess the financial burden represented by the children for the separated mother, for example, the authors apply the CU ratio linked with the children out of the total household CUs to the woman’s disposable income (salary minus the taxes paid, plus the benefits received and the support payment by her ex-partner for the two children in her care). But there is nothing to say that the separated mother does not allocate more resources to the children than is estimated by the CU ratio (with regard to housing, for example, she might sleep in the living room so that the kids each have their own room).

The methodological criticisms made of equivalence scales limit their use (see Martin and Périvier, 2015). They are not suitable for comparing the living standards of individuals, but only the living standards of households of different sizes.

What about the child’s standard of living?

There is not much literature estimating the standard of living of separated parents. To fix CUs per child in accordance with the marital status of their parents (in couples or separated), the authors rely on an Australian study that leads them to increase the CU attributed to children once the parents are separated. The cost of a child of separated parents is higher than that of a child living with both parents. They opt for the following formula:

– a child living with both parents corresponds to a CU of 0.3;

– a child living with the mother in conventional custodial care is 0.42 CU and 0.12 for the non-custodial father, i.e. 0.54 total CU for the two households.

Thus the cost of a child of a separated parent is 80% higher than that of a child living with both parents. It is likely that most separated parents do their best to keep the lives of their children unchanged after a separation. An approach that seeks to maintain the child’s standard of living makes it possible to take this into account. By increasing the cost of children by 80% when they live with both parents, and redistributing this in proportion to the CUs allocated for the children of separated parents, the custodial parent has a greater loss in living standard than that of the non-custodial parent (see the Table). This method is also questionable because it applies the additional CUs of children of separated parents over children living in couples to the monetary cost calculated in the case of a couple raising the children. But if this approach is chosen, then the result is reversed.

Any statistical analysis is based on assumptions used to “qualify” what we want to “quantify”, which is inevitable (either because we do not have the information, or for reasons of simplification and to facilitate interpretation). Assumptions that are too strong, results that are too sensitive, and perfectible methodologies are the daily lot of researchers. Providing insights, asking good questions, opening up new perspectives, feeding and feeding off of contradictions – this is their contribution to society.

The study published by France Strategy has the merit of initiating a debate on a complex subject that is challenging for our tax-benefit system. But the answers that it gives are not convincing. While the authors acknowledge that, “The interest of these simulations is above all illustrative,” they nevertheless also want that “at least they provide judges and parents with a tool to simulate the financial position of two households that have resulted from a separation by integrating the impact of the tax-benefit system”. This seems premature in view of the fragility of the results presented.

 

[1] To compare the standard of living of households of different sizes, equivalence scales are estimated from surveys and using a variety of methods. They are used to refer to an “adult equivalent” standard of living, or a “consumer unit” (CU). From this perspective, the standard of living of a household depends on its total income, but also on its size (number and age of its members).

[2]  While Figure 7 of the working document summarizes the situations by the number of children, in the note the focus is on the case with two children.

By Françoise Milewski and Hélène Périvier

Legislating to promote equality

The laws on equality in pay and in the workplace have come a long way since 1972, from the affirmation of the principle of equality to the production of a detailed numerical diagnosis that puts flesh on the bones of inequality (via the Comparative Situation Reports that have been drawn up since 1983 under the Roudy law) as well as to the duty to negotiate. The 2006 law paved the way for hitting recalcitrant companies with financial penalties, as set out in an article in the 2009 law on pensions. There were numerous attempts to limit the scope of the law up to 2012, when things were more or less clarified: companies are now obliged to produce a CSR, which reports annually on the state of inequality in well-defined areas; they must then conduct negotiations on occupational equality and equal pay and, if there is no agreement, they are required to take unilateral action. There are exhaustive controls, with agreements or plans to be filed with the government (no longer on a one-off basis as in the first formulations of the implementing decree). Companies that fail to comply with the law are put on notice to remedy this on pain of financial penalties of up to 1% of payroll.

The duty to negotiate entails collective management of the issue. Since 2012, the number of agreements signed has increased, as have formal notices and sanctions. While the content of the agreements and plans is often too general, it’s a start. The framework law of 4 August 2014 on equality has complemented and strengthened these arrangements.

Simplification: naïveté or retreat?

On the occasion of the Rebsamen bill on social dialogue, this long legislative process is suddenly being called into question under the pretext of simplification. In the bill’s initial version, the requirement to produce a detailed diagnosis in a CSR is gone, having melted into the company’s single database. The duty to negotiate on occupational equality also disappears, integrated into other negotiations (quality of life at work).

Given the extent of the reaction (associations, individuals, unions, researchers, etc.), the three ministries concerned issued a statement reaffirming certain principles, including that “it shall continue to be obligatory to transmit all the information that is currently found in the CSR”. Amendments will be tabled to that effect. But nothing is settled. The gender indicators remain integrated into the single database, so the CSR loses its specificity. Negotiations that focus on equality are not restored, and their frequency remains unclear (annual? triennial?). Uncertainty remains.

Whatever the outcome of the parliamentary debate that is starting up on social dialogue, business has been given the signal that equality policy can be challenged, that previous requirements are ultimately not all that imperative, and that the measures taken in recent years can be relativized in the name of simplification.

If, by leaving it up to the social partners to negotiate on gender equality, this issue had emerged on its own and led to significant progress, no law on the subject would have been necessary. It was in response to inertia and persistent inequality that constraints were imposed on companies. It is because our society needs to make gender equality a fundamental principle that laws, coupled with constraints, were approved. The complexity of the social dialogue on this subject reflects the resistance of the different parties. This simplification is at best naive, and at worst a refusal to come up with public policy to promote equality.

In the field of equality, vigilance is vital. Removing the constraints means going back on the principle of equality. A desire for equality requires clear, ongoing political will: continuity and coherence in public policy is crucial.

This is the meaning of a statement by men and women researchers that was published on the Les Echos website on 19 May.

 

By Hélène Périvier

The bill on equality between women and men, approved by the Senate on 18 September 2013, includes a component aimed at modifying the arrangements for access to the allocation of parental leave [1] by introducing what is called the free choice of activity (“CLCA”). The latest OFCE Note (no. 34 of 26 September 2013) analyzes the consequences of this measure for gender equality and proposes other possibilities for a broader reform.

The right to the allocation of parental leave is a family right: it is allocated to a parent who cuts their working time or ceases working altogether in order to care for a child, for a maximum period of 3 years. Noting that 98% of the beneficiaries are women, the law aims to encourage fathers to take it up: henceforth, out of the 36 months allocated for parental leave, 6 must be taken by the other parent. In other words, once the mother has taken 30 months of parental leave, the father must take over or else the family will lose the remaining 6 months. The UNAF, which opposes the reform, has published a survey on “fathers and parental leave” on its website. Arguing that the two sexes are complementary, it opposes the principle established in the law aimed at promoting the sharing of family responsibilities between mothers and fathers. Furthermore, the lack of childcare for young children is highlighted as a barrier to any modification of parental leave, on the grounds that this would accentuate the organizational constraints on parents of young children. Nevertheless, the gendered nature of parental leave is making this programme an obstacle to equality, even if some of the recipients say they use it out of personal choice. Making progress on gender equality thus requires reforming the mechanisms for access to parental leave. But will the proposed legislative changes be sufficient to shake up the boundaries of the existing sexual division of labour?

Redistributing the constraint between mothers and fathers

Given the struggle against the discrimination that affects most women, failure to make the CLCA reform would amount to introducing the freedom to use leave by some mothers and the freedom not to use it for all fathers. Parental leave is of course not the only factor responsible for gender inequality, but it is a driving force, and occupational inequalities in turn reinforce this inequality.

A policy designed to promote occupational equality cannot therefore avoid the reform of parental leave. Ending this vicious cycle necessitates major changes to this programme. Leave that is shorter and based on an individual right that is non-transferable between spouses, with compensation linked to the beneficiary’s income, would undoubtedly be more attractive to fathers and would promote equality (Méda and Périvier, 2007). While not directly egalitarian in itself, such a scheme would have the enormous advantage of ensuring women’s autonomy in relation to their spouse, thereby making economic empowerment a principle of public policy. But it is not possible to shorten the duration of parental leave without having first filled the gap in childcare for young children, which is currently estimated at 350,000 places [2]. The re-organization of leave should therefore be part of an overhaul of early childhood care. Otherwise, shortening parental leave would wind up further increasing the burden weighing on parents, and mothers in particular. An ambitious early childhood care policy, featuring short parental leave paid in proportion to salary, would promote equality. This would require significant public expenditure, about 5 billion euros a year (Périvier, 2012). The trade-offs being made in the course of the government’s budgetary adjustments point, however, to cutbacks in public spending.

In fact, due to a lack of funding, the proposed reform of the law is modest and will not really rebalance the sharing of family responsibilities between women and men. But it has the merit of highlighting the contradictions in society with respect to equality: without a requirement to share parental leave, this would be taken up only by women. The introduction of a period of parental leave allocated to the father will not directly increase the burden resulting from the shortage of childcare: the right to the allocation of parental leave is still 36 months for the family. It will merely spread the load between mothers and fathers. The trade-off facing fathers is the same as what mothers have faced for a long time. Given the low flat-rate amount of compensation, few fathers are likely to be tempted to take this leave. However, while the guidelines on budgetary matters are closing the door on any ambitious reform of early childhood care, women must not be the only ones to bear the consequences.

Reforming parental leave is thus imperative for equality.

[1] It is important to distinguish the allocation of parental leave  as such from parental leave in terms of labour law (Labour Code Article L. 122-28-1), which, subject to certain conditions, guarantees that all employees will regain their job after taking parental leave for a period of one year, which is renewable three times. The first is paid by the CAF within the broader context of family policy, subject to certain conditions (rank of the child, past activity, etc.). The conditions of access in terms of past activity are more flexible for granting eligibility for the allocation than parental leave in the strict sense. In fact, only 60% of CLCA recipients benefit from a guarantee of re-employment (Legendre and Vanovermeir, 2011).

[2] See, in particular, the Tabarot Report, Périvier 2012.