1Most OECD countries have observed an increase in divorce rates since the early 1970s. Between 1970 and the end of the 2000s, the total divorce rate rose from 1.0 to 1.9 per 1,000 in the European Union and from 0.8 to 2.0 per 1,000 in France.  This evolution is linked to changes in social norms and conceptions of marriage (seen less as a religious commitment and more as a partnership), which tend to trivialize divorce. The demographic trend has been accompanied by changes in the law that make divorce easier, including when requested by only one of the spouses. In France, the Law of 11 July 1975 further facilitated access to divorce by authorizing it on grounds other than fault. More recently, the Law of 26 May 2004 set out four divorce procedures: divorce by mutual consent, “accepted divorce”, irretrievable marriage breakdown, and divorce on the grounds of fault. The main objective is to simplify divorce procedures and settle the financial arrangements more quickly. Globally speaking, couples marrying today are more likely to divorce than previous generations owing to greater social acceptance of divorce and simpler divorce procedures.
2This shift has been the focus of extensive research in the social sciences. On the basis of the typology proposed by Lambert (2009) to classify sociological research on divorce, two types of research can be distinguished. The first concentrates on the causes of divorce, the second on the consequences for former spouses and their children. The same distinction applies in economics. In the economic analysis of law, a number of theoretical and empirical studies have sought to determine whether or not the introduction of no-fault divorce has led to an increase in the divorce rate. In parallel, in the field of the economics of social policies, considerable research has focused on the impact of separation on changes in the standard of living of former spouses and their children and, to a lesser extent, on the impact of separation on the labour market activity of the former spouses. This research shows that divorcing spouses generally see a drop in their standard of living due to a reduction in economies of scale. While the extent of this drop varies from one country to the next (Aasve et al., 2007; Andress et al., 2009), several facts can be established. In particular, the standard of living decreases more often for women than for men following a divorce (Bonnet et al., 2015), especially if the woman has custody of the children. In addition, the risk of poverty following a divorce is higher for women than for men (Ananat and Michaels, 2008). Furthermore, this same literature shows that public and private transfers between former spouses (child support or alimony) help to reduce the risk of poverty and to narrow the gender gap in living standards following a divorce (Bratberg and Tjøtta, 2008; Poortman, 2000). In other words, divorce has an economic cost for the spouses. This cost is shared – though not always equally – between the former spouses, but also between the two spouses and public solidarity.
3The aim of our article is to examine these costs by considering divorce as a risk in the economic sense of the term, i.e. as the combination of a probability of occurrence and a level of loss.  With divorce now socially accepted, the aim is not to identify ways of reducing the likelihood of divorce, for example by devising incentive mechanisms to prevent its occurrence or by reforming divorce law. The article focuses entirely on the coverage instruments available for protecting individuals against the financial losses arising from divorce. 
4The economic consequences of a separation are not specific to divorce alone; they concern all types of separations. But the extent of coverage instruments depends on the type of union (marriage, civil partnership, consensual union). Marriage is the form of union with the broadest range of coverage schemes, notably owing to the obligations between spouses created by marriage and the procedures for sharing the couple’s assets through the marital property regime. Consequently, by focusing analysis on the separation of married couples, the article covers the broadest possible range of instruments covering the risk of separation. That said, the analyses developed in this article remain relevant for all other types of union, providing that the coverage instrument is accessible. 
5This article makes a twofold contribution. First of all, it provides a complement to sociological studies of the aftermath of divorce. As stressed by Lambert (2009), sociological research on divorce, in France at least, tends to focus mainly on parenting issues. But we believe that post-divorce conjugal issues also deserve to be examined. Secondly, by defining divorce as an economic risk, the article provides justification, at least in part, for the roles that can be attributed respectively to the family, the state and the market in the coverage of divorce costs.
6The article is in three parts. Section I presents an economic discussion of the nature of divorce risk, particularly in financial terms. The aim is to demonstrate that the economic costs of divorce are not of an exclusively private nature; they also have a social dimension in that the consequences of divorce are not borne solely by the ex-spouses. The answer to this question determines, from an economic standpoint, the different ways of covering divorce costs. In Sections II and III, the article addresses the principles and the various approaches applied for covering the economic costs of divorce.
I – Divorce as a risk with costly consequences at both private and collective levels
7Divorce risk has a dual, private and public dimension. Divorce can be considered as a private risk in that it impacts the wealth of the ex-spouses. But it can also be seen as having a public dimension since its consequences affect society as a whole.
1 – Divorce as a private risk
8The realization of divorce risk generates two types of cost for the ex-spouses. First of all, divorce results in direct costs in the short term. In addition to the expenses linked to the procedure itself,  these costs mainly correspond to the overall decrease in the ex-spouses’ standard of living due to a loss of economies of scale after separation. This decrease is not generally distributed evenly between the former spouses. The main cause for this imbalance lies in each individual’s own income. When one of the spouses accounts for the majority of the couple’s income, his or her standard of living increases substantially following the divorce while that of the other spouse decreases. The second cause of asymmetry is linked to the presence of children and the fact that their main residence is at the home of one of the former spouses. This imbalance, along with the overall decrease in standard of living, may also push one of the two parties into poverty.
9Alongside these direct costs, divorce also generates indirect costs, resulting from decisions taken during the marriage, notably concerning marital investments. The costs in question stem, in part at least, from the fact that while relations within the couple were initially organized in a long-term perspective, the decision to divorce necessarily shortens the time horizon. The problem is posed in particular when one of the spouses has made specific long-term investments in the marriage and the divorce deprives him or her of the returns (Cohen, 1987). The economic literature provides concepts that are useful for analysing the nature of these investments and their consequences. For economists of the family, marriage is a long-term contractual relation governed by law (Cigno, 1991). Divorce marks the termination of this contract, amicable or otherwise. From this viewpoint, marriage is at the root of a major problem if one of the spouses (say, partner A) has made specific investments in the domestic sphere (by caring for and raising the children or managing the household, by financing the other spouse’s education or early working career). These investments bear fruit in the medium or long term. Partner A (most often the woman) profits from the marriage partly at the time he or she makes the investment but mostly once the investment is completed. In contrast, the other spouse (say, partner B) profits from the marriage immediately but pays the costs only in the long term, when partner A’s yield decreases sharply (the children have grown up, the career of partner B who has invested in the job market no longer depends on partner A’s investment, and so on) while B continues to pay for A’s upkeep. So for partner A, divorce deprives him or her of a return on investment, partner B having profited from partner A’s investment without paying the long-term cost. Furthermore, domestic investments are hard to transfer to other domains. The skills acquired by partner A who devoted part or all of his or her time to domestic life are of little value on the job market.  As a result, divorce cancels out the value of the investment made during marriage.
10The indirect costs related to specific investments can be assessed not only in the short term but also in the medium or long term. Where these investments have kept the spouse away from the job market, it is more difficult for them to return after a divorce owing to their low employability (initial career interrupted by periods of inactivity). The spouse’s employability level is further reduced if they are a lone parent, as they are less available for work owing to the presence of children. In addition, the consequences of the specific investments are also felt when a divorce occurs later in life. Irregular or less intensive presence on the job market (due to the presence of children) affects the replacement incomes received after retirement (Bonnet and Hourriez, 2012).
2 – Divorce as a risk of a social nature
11Divorce also generates social externalities in that its impact extends beyond the spouses themselves, affecting any children they may have and, potentially, society. First of all, divorce may be a cause of lone parenthood, which is known to increase the risk of poverty for children. As demonstrated by Legendre (2003), the poverty risk for a child living with a lone parent is twice as high as that of a child living with two parents. And child poverty comes at a cost for the community. In the short term it increases social expenditure on means-tested family benefits (RSA welfare benefit, family allowances). Above all, in the medium term, child poverty compromises the human capital of future citizens. A number of studies, mainly British and American, have shown that poor children are more at risk than others of experiencing adverse outcomes over their lifetimes, such as poor performance in school (CERC, 2004), difficulties in entering the job market, poverty as adults (Rodgers, 1995), and low income or part-time working (Duncan et al., 2010).
12Divorce also contributes to the development of gender-based economic inequalities. As demonstrated by extensive statistical research in the United States and Europe, women’s standard of living after a separation drops sharply on average, while separation has less of an impact on ex-husbands. Inequalities in post-divorce standards of living, resulting in part from the choices made before the divorce, reduce the woman’s ability to generate income. In the large majority of cases, it is the woman who adjusts her working career to take care of the couple’s children (Pailhé and Solaz, 2006). Also, in the event of extended career interruptions, women’s wages tend to be lower when they return to work, and remain so over the long term (Lequien, 2012). In other words, domestic specialization is painless for the wife during the marriage because resources are pooled, but becomes costly after a divorce, in both the short and long term. Further, domestic inequality often continues after the marriage, given that only a minority of divorced couples have alternate custody of their children and that the children’s main residence is most often at their mother’s home.  Globally, then, lower income and a greater number of consumption units for the ex-wife than for the ex-husband add up to a more disadvantageous economic post-divorce situation for women than for men. This asymmetric financial impact may be lessened for women if they form a new couple. But the attractivity of divorced men and women on the marriage market is not the same, with the former remarrying sooner and more frequently than the latter (Cassan et al., 2001).
II – Private divorce risk coverage
13From the standpoint of economic analysis, divorce risk should, at least in part, be covered privately. Given that divorce is a specific risk – its occurrence depending on the decision of at least one of the spouses – one may consider that the ex-spouses should be responsible for the consequences. Economic analysis can be used to explore the ways in which such private coverage could work.
1 – The theoretical foundations of private divorce risk coverage
14On the basis of the economic analysis of risk, two reasons may be put forward to justify the coverage of divorce risks by the individuals themselves, given the responsibility involved in deciding to make the risk occur, and the possibility of taking action to prevent it from occurring.
15The realization of the risk results from the decision of at least one of the spouses to put an end to the union. Unlike other types of risk, divorce is a risk with an endogenous probability of occurring, as its realization always depends on the decision of at least one of the spouses rather than an agent external to the couple.  This affirmation may be nuanced, in particular where the request for a divorce comes from just one of the spouses. For the person subjected to the decision of the other, divorce appears to be an entirely random event independent of their will. Even so, the likelihood of divorce is not necessarily exogenous since the decision to divorce by the other member of the couple may result from incautious behaviour on the part of the spouse not asking for a divorce. Consequently, the risk is realized either by the decision of one of the spouses (unilateral divorce), and in this case the non-deciding partner bears the brunt of the risk, or by a joint decision to put an end to the union (bilateral divorce). That being so, the partners may be held “responsible” identically for the occurrence of the risk, whether they make a joint decision to divorce or only one of them makes that decision.
16Another justification for private divorce risk coverage could be that the spouses are in a position to prevent the risk from occurring. They can act upon the probability of divorce by being faithful and attentive to the other, for example, or by investing in the household (which, in the terms of Ehrlich and Becker (1972), is a form of self-protection). We may thus consider that if a divorce occurs it is because the spouses have not made sufficient effort to maintain their union and that this responsibility is shared, without necessarily being identical. Beyond the actual decision to divorce, the chosen legal procedure (divorce by mutual consent, on grounds of irretrievable marriage breakdown or fault) could thus serve as a way of dividing responsibilities and, by so doing, of sharing the costs of divorce between the two ex-spouses.
2 – Possible ways of handling divorce as a private risk
17A number of instruments exist for covering divorce costs, but they are not without their limits.
The diversity of coverage instruments for private risk
18The majority of industrialized countries have a legal form of risk coverage for divorce, known variously as alimony, maintenance or spousal support. Awarded to spouses whose standard of living has been compromised, alimony takes the shape of a private transfer from the wealthier to the less wealthy spouse. For the latter, it constitutes a form of ex post coverage (after the damage occurs) of divorce risk.
19From the standpoint of economic analysis, such a transfer is justified by a criterion of effectiveness, as alimony encourages spouses to specialize in optimal sharing of tasks by ensuring that the spouse who invests in the family receives compensation for that investment in the event of a separation. In the founding model of the economic analysis of alimony, Landes (1978) shows that its role is to compensate the wife for the opportunity costs incurred when marrying (notably giving up on her own career), and that by doing so it enables the spouses to attain optimal output by encouraging each of them to specialize in the task in which they are the most productive. To justify alimony, Cohen (1987) insists on the specific – and thus hard to redeploy – nature of these domestic investments, which, in the event of divorce, cannot be recouped by the person who made them. The specific investments approach, inspired by the economic analysis of contracts, can be used to specify the nature of the loss to be offset. According to the typology proposed by Shavell (1980) to characterize the various types of loss resulting from breaches of contract, alimony may be considered, as did Bolin (1994), as a means to cover three types of loss. The first is the loss resulting from having furthered the career of one’s spouse without receiving any gains from that career owing to the separation. In this case, alimony ensures that in the event of separation, the wife is rewarded for her investment in her husband’s professional advancement. To calculate the loss to be offset, we must consider the husband as if he had never been married and determine the value of the additional human capital obtained through the marriage. The second type of loss may correspond to one or other spouse giving up their professional career to make a greater investment in their family. To determine the value of the loss to be offset, the “wronged” spouse has to be considered in terms of the situation they would have been in had they not married, and the present value of the income shortfall from not having had a full-time and continuous professional career must be calculated. The third type of loss may include all the material and affective losses stemming from the divorce. Here, the value of the loss is calculated in terms of the post-divorce decrease in well-being. Calculating the loss involves considering what the spouse’s situation would have been if he or she had remained married. It should be noted that, from an economic standpoint, nothing justifies treating the consequences of separation differently for different types of union. But in France, alimony exists under law for married couples only.
20Alongside alimony, ex ante instruments (before the risk is realized) are, or may be, used by the partners to protect themselves against the economic costs of divorce.
21A basic form of ex ante coverage consists in both spouses building up precautionary savings or limiting their specific investments in domestic life and having a professional activity throughout the marriage, the risk being, under the economic approach of alimony, that the couple’s output is not maximized.
22The spouses may also seek to self-insure against divorce risk by signing a marriage contract that allows them to anticipate ex ante the economic consequences of a divorce. The marriage contract is analysed by economists in the same way as alimony. Based on contract economics and on a vision of marriage as enabling specific investments aimed at optimal domestic output, the contract is seen as a tool that also serves, by securing specific investments, to maximize marriage gains (Cigno, 1991). More specifically, by providing for marriage assets to be shared in a way that favours the partner who makes the specific investment, the contract creates “positive” incentives. The question here is to identify the optimal rule for sharing out the couple’s assets at the moment of the divorce, i.e. a distribution that will generate the best incentives for investing in the marriage and, by doing so, maximize the gains (Rainer, 2007). Economists stress the potential advantages of prenuptial contracts over alimony or the rules set out in statutory property regimes, insisting on the efficacy of giving contractual autonomy to the spouses (Smith, 2003).
23A third form of the ex ante coverage of divorce risk could be envisaged in which the spouses transfer the coverage to a third party, providing that the conditions applicable to private insurance are fulfilled.  Households could take out an insurance policy at the start of the marriage to guarantee compensation for the spouse most financially disadvantaged by the separation. The amount of the premium would be proportional to the value of the potential loss for the spouse in question and could be reassessed in line with the professional and family choices made by the household throughout the marriage.  Each spouse could also take out an insurance policy in their own name to provide them with a certain replacement income in the event of divorce  or, at the very least, coverage for some of the costs stemming from the separation (and notably the costs of proceedings).
The limits of divorce risk coverage instruments
24The individual precautionary strategy is neither a universal nor certain coverage instrument. Precautionary savings are a form of coverage reserved for individuals able to put money aside, excluding those at the bottom of the earnings scale. Also, in the long term, nothing guarantees the value of the capital saved by individuals unless they diversify their savings sufficiently. Likewise, the quality of the coverage obtained by remaining in the job market depends on the individuals’ educational level and their employment situation. If a person is on low wages or if the unemployment rate is high, then coverage will be poor.
25Concerning coverage under a collective system of risk mutualization based on a private insurance mechanism, the generosity of the scheme will depend on the amount of the premiums paid by the spouses during the marriage.
26Alimony raises the question of its predictability, at least in cases of contested divorce.  Because the law leaves room for interpretation in the granting and setting of the amount,  alimony does not provide automatic coverage in the event of a divorce.  According to the findings of Doriat-Duban and Bourreau-Dubois (2013), French judges set thresholds on eligibility for alimony (in terms of the degree of income disparity between the partners and the length of the marriage).  In other words, as the activation of solidarity between ex-spouses is subject to the discretion of a judge, the party with the sharpest reduction in their standard of living is not certain to receive compensation, and if compensation is granted, the amount received may not be equal to the magnitude of the loss.  Alimony, then, is a partial and non-systematic (and even random) coverage instrument.
27In contrast to alimony, marriage contracts provide the partners with predictability on how the costs of a divorce will be divided. But another question is raised, as it is not certain that the marriage contract protects the party for whom the divorce comes at the greater cost. As highlighted by Oosterbeek et al. (2003), this is because the specific domestic investments restrict the exit options of the specialized partner and thus limit his or her power of negotiation. It is for the same reason that Cigno (1991) and Smith (2003) recommend signing a marriage contract at the start of the marriage when the transaction costs related to negotiations are not yet too high. The uncertainties over the degree of protection provided by the marriage contract becomes manifest when the various matrimonial regimes are scrutinized. Not all contracts have the same effects in the event of a divorce. Community of property ensures an even division of the couple’s joint assets after marriage dissolution, while other regimes protect the wealthier spouse and may lead to increased divorce costs for the more vulnerable one (McLellan, 1996). This is the case under the separate property contracts, which protect the wealthier spouse by allowing them to keep all the assets acquired before and during the marriage and by limiting the sharing of assets to those under undivided ownership. Moreover, the current trend in the event of separation appears to be towards a reduction in solidarity between partners, as demonstrated in France by an increase in separate property contracts among married couples (Frémeaux and Leturcq, 2013).
28Beyond the type of contract chosen, it may become impossible to negotiate a marriage contract for various reasons. First, the parties may be affected by optimism bias (Smith, 2003), particularly in first marriages. For example, an American study (Baker and Emery, 1993) based on a survey of future newlyweds, shows that they have a fairly accurate estimation of the overall risk of divorce but strongly underestimate their own risk of divorcing. The authors ascribe this optimism to a representativeness bias (Kahneman and Tversky, 1982), with the interviewees considering themselves as not representative of the population as a whole. Even in the absence of such bias, other sources of excess optimism also exist, as highlighted by Smith (2003). For example, the parties may also underestimate the risk of divorce. Consistent with the work of Kahneman and Tversky (1979), individuals tend to underestimate the high probabilities of negative outcomes, and this reduces their incentive to protect against divorce risk. Second, the idea of negotiating a marriage contract may be perceived as a negative message as to the solidity of the marriage, thus making such a negotiation rather unlikely. These same limits may be at the root of under-insurance against divorce risk.
29Their differences aside, the marriage contract and alimony both share the same limits in terms of divorce risk coverage for the party the most penalized by the marriage breakdown. First, the two systems reduce the costs of divorce only if there is an income to be redistributed or wealth to be shared. In other words, this type of coverage does not work systematically. Second, both alimony and the marriage contract may place upward pressure on divorce costs but without necessarily being able to cover them. By guaranteeing compensation in the event of separation, these systems encourage specific investments (Bianchi et al., 2014). If the specialization is strong, for the spouse who specializes in the domestic sphere it may entail a substantial reduction in earning power over the short and long term, which automatically increases the extent of his or her losses in the event of separation. Yet alimony and marriage contracts are designed to help people cope with the short-term costs of a divorce rather than compensate for the longer-term losses resulting from separation.  The fact that the very existence of compensation is liable to increase the magnitude of loss is a well-known mechanism in the economics of insurance, one referred to as a problem of moral hazard. 
30In the light of this overview, three main coverage approaches may be distinguished. The first is based on inter-individual solidarity between spouses (alimony and marriage contract), the second on the pooling of risks (private insurance) and the third on individual coverage (remaining in employment, precautionary savings). The first corresponds to the systems enshrined in law. Having existed for a long time in most countries, these systems apply to married couples only. The second, based on market instruments, appears to be emerging in some countries to specifically cover certain types of costs stemming from a divorce (mainly the costs of proceedings, and more rarely the payment of child support). The third approach is also based on a market logic, that of the job market and the financial markets. According to results obtained from US data, some of these different forms of coverage can be partly substituted one for the other. For example, it appears that in a context of no-fault divorce, women living in US states where the law provides for equal sharing of the couple’s assets tend to work less than women living in states where the law is less favourable to women who divorce (Chiappori et al., 2002; Voena, 2012).
III – The role of the state in divorce risk management
31From the standpoint of economic analysis, public intervention in the management of divorce risk, in both its social and private dimensions, may be justified. This intervention may be direct or indirect.
1 – The theoretical foundations of public intervention
32Divorce risk is not considered as a social risk in that the material consequences of divorce do not give rise to coverage by a social protection system,  notably a public one. However, viewing divorce risk as a kind of social risk could be justified under Pollak’s interpretation of the question (2011). For Pollak, a risk may justify collective coverage if it meets four criteria: it has material consequences, it is probabilizable, there is adverse selection  and there is no legal fault. The first criterion is respected where the divorce affects the economic situation of the ex-spouses, notably through an exceptional increase in their expenditure. The second criterion relative to the probabilizable nature of the risk is also fulfilled in that it is easy to calculate the probability of a divorce based on the frequency of divorces observed in the population (and notably according to the length of the marriage).  In addition, individuals are not equal in the face of divorce risk,  even if they are all likely to be exposed to divorce once married, so a risk of adverse selection exists. Lastly, divorce is now largely dissociated from any reference to a fault or responsibility, except in the case of at-fault divorce. But divorce risk is not considered as necessarily deserving of public coverage, as society does not support recognition of collective responsibility in this case. The main reason is undoubtedly the fact that, basically, society considers that the individuals concerned are responsible for divorce risk (Pollak, 2011) and that divorce is not a general dynamic of instability in the family structure to which individuals are subjected, and against which they are powerless. In other words, social protection should not be required to compensate for an increase in expenditure subsequent to the choice of a more expensive lifestyle. It may be considered that the costs borne subsequent to a divorce are offset by the greater well-being resulting from separation, for one of the ex-spouses at least (De Singly, 2011). Moreover, as social inequalities in divorce risk do not give rise to clearly anti-redistributive mechanisms,  there is not necessarily a call for society to reduce them. In other words, while there is indeed a risk of adverse selection, it is not clear that it works to the detriment of the most disadvantaged categories. For all these reasons, there is no basis for designating divorce as a fifth risk, as was envisaged for old-age dependency, for example.
33While divorce risk may not call for direct coverage by the public authorities, two reasons can be identified that may justify public intervention in the management of divorce risk.
34The first is based on the negative social externalities generated by divorce, to the detriment of children and of women. Child poverty should be addressed by the public authorities for reasons of fairness and effectiveness alike. First of all, as stated by the French Council for Employment, Income and Social Cohesion (CERC, 2004), “in the matter of social justice, it is the duty of society to redress the inequalities suffered by people who are not responsible for the situation in which they find themselves. This applies to children more than to any other person”. Secondly, as stressed earlier, child poverty erodes the value of the human capital of a part of the population. So it is in the interest of the public authorities to protect the interests of children living in lone-parent families when these are synonymous with poverty. Regarding women, and as pointed out above, divorce tends to have more adverse financial consequences than it does for men. Here again, one may consider that, in the name of a criterion of social justice relative to gender equality, it is the state’s role to intervene in order to limit such inequality following a divorce.
35The second reason for public intervention may lie in the desire to regulate the private management of the financial consequences of divorce wherever they lead to unfair situations for one of the parties or to situations of inefficiency, where the individuals concerned are not covered for a risk that could be insured against. From a purely couple-centric viewpoint, the spouses are not always mindful of the risk. As long as the marriage “works”, divorce is a random event for both spouses who know neither if nor when they will separate. They may suffer from a certain short-sightedness or even a cognitive bias (see above for optimism bias) that causes them to minimize the actual risk of divorce. They may also imprecisely assess the financial and human consequences, as the extent of the losses also depends on the occupational choices of both partners and their ability to “bounce back”, including on the marriage market. In other words, the parties are not fully qualified to evaluate the costs of divorce and plan for such an eventuality, or to prepare for the related social costs. Moreover, power relations during a divorce may be unequal, leading to an unfair division of assets between the partners. Consequently, there are grounds for considering the regulation of divorce risk at a more collective level.
2 – Public systems for managing divorce risk
36The state can implement social and tax policies and wield its legislative power, both to limit the social costs of divorce and to improve private management.
Systems for limiting the social costs of divorce
37While the community is not disposed to providing social insurance against divorce risk (see above), it does consider as a new social risk one of the consequences of separation, namely situations of lone parenting (Eydoux and Letablier, 2009; Pollak, 2011). Lone parents are covered by parenting support policies, either in the shape of targeted services (increased RSA welfare benefit) or socio-fiscal transfers aimed at financially supporting households (lone-parent or not) with children.  The introduction of guidelines setting the amount of child support for the children of separated parents may also be considered as another way for the legislator to directly regulate the economic consequences of divorce, with a view to limiting its negative impacts on children. This is notably the case when, as in the United States, the judges tasked with resolving differences between the parents are obliged (barring exceptional waivers) to award the amount set by the guideline. In other countries, including France, guidelines can be used, at the very least, to guide the judges’ decisions. The introduction of a rule for calculating the amount of child support subsequent to a divorce provides the state with a way of ensuring the material well-being of children with divorced parents by specifying the parents’ support obligations. Because the calculation rule is based on objective estimates of child costs, it guarantees an adequate standard of living for the child, i.e. consistent with the parents’ abilities to contribute (Bourreau-Dubois and Jeandidier, 2013), and favours better future compliance with judicial decisions.
38The gender inequalities that emerge at the time of divorce are also not considered to constitute a social risk. But these inequalities are taken into consideration, to varying extents, by social protection law. In some countries, “pension splitting” applies in the event of divorce. The principle consists in grouping the pension entitlements acquired by the man and the woman during their union and then splitting them evenly between the two at the moment of the divorce. This rule applies in Germany, Switzerland and Canada.  In the UK, pension entitlements are included in the overall assets of the household, which are then shared between the two ex-spouses. This system makes up for the low individual entitlements of women who reduced their working activity during the marriage to make specific investments in domestic life. France has for many years operated a system of family-related pension entitlements designed, notably, to offset the impact of career interruptions on the pension levels of mothers (Bonnet and Hourriez, 2012). Last, gender inequalities subsequent to divorce are, as a last resort, covered by social welfare systems. In France, RSA welfare benefit and the minimum pension form an ultimate safety net for single women in situations of poverty.
39Alongside these systems providing ex post compensation for the gender inequalities resulting from divorce, the state may also use tools aimed at limiting the occurrence of gender inequalities during the marriage which, through a knock-on effect, helps to reduce the risk of poverty post-divorce. The objective here is not for the state to remedy difficult situations or redress gender inequalities but to change the behaviour of individuals. As seen earlier, specific investments in the domestic sphere are very costly for women if they separate. These investments reduce their earned income and, if they have custody of the children, put them at a high risk of poverty. Generally speaking, then, the idea is to encourage mothers to remain in employment after the birth of their children. The trend today is towards greater intervention in this respect on the part of public authorities, as demonstrated by the recent reform of parental leave in France, voted as part of the law on gender equality that, by encouraging fathers to take parental leave, aims to limit gender inequalities associated with divorce. Further upstream in the cycle, the state may also seek to improve the educational levels of women (Guvenen and Rendall, 2013). Lastly, it may use tax incentives, by reducing taxation of dual-earner households, for example, through individualization of income tax (Esping-Andersen, 2008; Landais et al., 2011).  Overall, the public interventions that can be rolled out to regulate divorce risk belong both to the traditional conception of social policy (the fight against poverty of lone-parent families) and the modern conception of social policy (aimed at fostering women’s employment) defended by Esping Andersen (2008), for whom the welfare state should be an investor rather than a nursemaid.
Tools for improving the private management of divorce
40The state may also improve the private management of the consequences of divorce by harnessing its power to state the law. If one of the spouses makes specific investments in domestic activity, this leads to problems of efficiency and fairness, as the divorce cancels out the value of the specific investment for the person who made it, and these investments lead to a drop in the standard of living that is not shared equally by the two ex-spouses. And yet it is not unreasonable to think that in a situation of unilateral divorce, the spouse who did not make the specific investment is unlikely to compensate their ex-partner without any prompting. By entrusting judges with the task of deciding on alimony, the state exerts control over the result of the negotiations between spouses in terms of divorce cost sharing. The objective is for the judge to ensure that the interests of both parties are respected and that the divorce does not generate an uneven distribution of divorce costs. The state may, as is the case in France,  leave it to the discretion of its agent – the judge – to choose the social justice criterion to be applied. In this case, it is up to the judge to decide on the threshold beyond which the inequality becomes unfair. The level of that threshold depends on the judge’s conception of solidarity between former spouses.
41The choice of statutory matrimonial regime is also an indirect way for the state to play a role in private divorce regulation. By designating community of property as the default legal regime, as is the case in France and Luxembourg, the law helps to protect the weaker partner by ensuring that they obtain half of the assets acquired during the marriage. In contrast, countries such as Austria and the United Kingdom, having opted for a separate property default regime (Granet and Dandoy, 2016), are, in principle, less protective of the spouse who made a specific investment and, by doing so, acquired fewer assets. Naturally, the spouses can sign a marriage contract different to the default regime, enabling them to protect the weaker party by planning for the fair distribution of property in the event of a separation. But owing to status quo bias, the choice of regime is not a neutral one. Because it is reasonable to assume that the spouses opt most often for the default regime and remain under that regime until a potential divorce, it would be in the state’s interest to choose the community regime as the default regime so as to protect the weaker partner in divorce.
42Finally, and again with regard to its legal power, one could imagine a system in which the state requires households to take out insurance policies to resolve a market failing that stems from a phenomenon of under-insurance. Given the existence of optimism bias or a preference for the present, the risk of underinsurance exists, with newlyweds potentially liable to underestimate their divorce risk. The mandatory nature of the insurance policy could be justified for other reasons, too, such as ensuring the solvency of the payer in the event of income loss due to an unexpected life event. 
43Initiated by at least one of the spouses, divorce is a risk (in the economic sense of the term) of a private nature, as it affects the wealth of the ex-spouses. One could thus argue that it is their duty to bear the associated costs. They have a number of coverage instruments at their disposal. Conventional instruments are based on private solidarity between the two ex-spouses or on an individual precautionary strategy consisting in not leaving the job market and in building up savings for a possible future divorce. Other coverage instruments, used much less often, are based on mutualized coverage via private insurance. But for reasons of efficiency and, above all, fairness, it is economically justified for the state to intervene in the private regulation of divorce risk. The law helps to ensure that divorce costs are shared in a relatively efficient and equitable manner by ruling on the payment of alimony or by establishing a default regime of community of property. And while divorce may be a private risk, it is also the cause of externalities that come at a social cost, including the heightened risk of poverty, stemming in particular from lone parenthood, and gender inequalities. It is the role of the welfare state to manage these two types of externality through direct public intervention in the form of policies to prevent lone-parent families falling into poverty, solidarity mechanisms in the pension system, and policies aimed at boosting women’s employment rate. In all, three actors contribute to the coverage of divorce risk: the family, the state and the market. Their intervention takes different forms and their respective roles in divorce risk coverage vary according to the financial situation of the ex-spouses and their ability to bounce back after a separation.
44This analysis could be extended by studying the degree of coverage of former partners by type of union, as some coverage instruments are not available to unmarried couples (including marriage contracts and alimony). It would also be interesting to analyse the way in which the respective interventions of the family, the state and the market are organized in different countries. For example, in Italy, family solidarity is a key form of support in the event of divorce,  while in the United Kingdom, divorce risk management appears to be handled at a more individual level, through an increase in women’s employment rate,  in an environment where the default matrimonial regime is that of the separation of property. For purposes of international comparison, a typology of the various forms of divorce risk coverage would be needed in order to compare cross-country differences in the economic consequences of separation.
Professor of Economics, BETA/UMR CNRS 7522, Université de Lorraine.
Correspondence: BETA, Université de Lorraine, 13 place Carnot, CO n°26, 54035 Nancy Cedex, email: firstname.lastname@example.org
A broader analysis of divorce in terms of well-being would involve a consideration of how divorce can foster greater well-being for one of the spouses, by freeing them from conjugal conflicts for example (Martin, 2007), or by enabling them to recover their identity (de Singly, 2011).
Non-divorce may also generate costs, for one of the spouses at least, as well as for the children (Martin, 2007), for example in the event of domestic violence or conflict. In these situations, while divorce may be preferable to continuing the union because it is less costly, the costs of divorce nevertheless remain, and may justify public coverage, which for the state entails paying social support to make the separation possible.
For example, only people who have been married are legally entitled to alimony in most countries. However, support for single parenthood is provided regardless of the person’s conjugal status.
Lawyers’ fees vary depending on the profile and location (Paris or province) of the law firm, the complexity of the case, and the contentiousness of the divorce. In France, according to the High Family Council (Haut Conseil de la Famille, 2014), the lowest fees are roughly €600 for the couple, rising to around €12,000 for contentious divorces. These costs may be reduced if the spouses receive legal aid. In addition to lawyers’ fees, the spouses have to pay notary fees if the couple’s property is divided between them.
The problem of employability is particularly acute for divorced female homemakers, whose skills are of little value on the job market (Parkman, 2001), as employers find it difficult to assess their human capital for tasks other than domestic ones (Bergmann, 1981).
According to a survey in 2012 by the French Ministry of Justice, in the event of divorce, residence at the mother’s home is pronounced in 69% of the decision made by the family affairs judge, alternate residence in 21% of cases, and residence at the father’s home in 6% of cases, with the remaining 3% of cases categorized as “other” (Carrasco and Dufour, 2015).
For example, divorce is different from the risk of loss in a car accident. In this last case, even if the individual decides to take their car or not, the accident remains an event independent of their will. The probability of an accident is exogenous. (With a divorce, the event is uncertain so long as the spouses do not know if the other will ask for a divorce, but becomes certain once one of the parties decides to make it happen.)
For a discussion of the possibility of developing private insurance against divorce risk, see Bourreau-Dubois and Doriat-Duban (2015).
While this type of insurance currently does not exist, it should be noted that the legal specialist Carbonnier mentions the possibility in his commentary on the law passed in France in 1975 on the creation of spousal alimony or “compensatory allowance”. Discussing the potential difficulties that payers of this alimony may have in honouring their payments, Carbonnier stresses that households could, in the future, count on insurance to “safeguard payment of allowances… Built around a life insurance policy in favour of the spouse, it can be seen as a marriage insurance policy serving multiple purposes” (Carbonnier, 2005 edition, pp. 1369-1370).
To our knowledge, thus far only a single attempt has been made to create this type of insurance. The project, initiated in 2010 by a US firm but not implemented, consists in selling the insurance policies to cover a drop in the standard of living following a divorce. The coverage takes the form of a lump-sum payment if the risk is realized, the amount of which depends on the premiums paid (very high). There is a waiting period of four years before receipt of payment (http://www.safeguardguaranty.com/What_We_Do.html).
This question does not arise for mutual-consent divorces in which the parties agree both on the principle and consequences of the divorce. There is no unpredictability, therefore, the partners having negotiated the payment of alimony and its amount under the shadow of the law.
In France, the Code Civil provides for alimony to be paid where a divorce engenders substantial disparities in the living conditions of the ex-spouses (Article 270) and provides the judge with the criteria for determining whether such is the case and for setting the amount of alimony (Article 271). The judge does not have an official guideline for setting this amount.
In France, alimony is awarded in only one divorce in eight (Roumiguières, 2004).
Based on the analysis of several hundred decisions by a provincial French appeals court, Doriat-Duban and Bourreau-Dubois (2013) show that the eligibility thresholds are 11 years of marriage and a €900 difference in income between the ex-spouses.
The amount of alimony set by the judge takes account of the difference in income between the spouses as well as the resources of the spouse paying the alimony. If the payer’s resources are limited, the amount of alimony is revised downwards.
For example, compensation in the form of a life annuity has practically disappeared in France, to the benefit of an endowment in the form of capital or a temporary annuity.
In this regard, a private insurance policy against divorce risk could limit the extent of the problem by including a clause in the policy limiting domestic specialization.
According to Dupeyroux (1998), “under the economic approach, too, the idea of social risk is not objectivized on the basis of risk characteristics, as no risk is social in itself (…), social risks are by their nature economic risks (as they have economic consequences on the income and expenditure of the agents) and are ‘social’ only in that a collective guarantee is established”.
In insurance, adverse selection refers to the fact that it is individuals with high risk levels who take out policies, thereby penalizing the insurer.
Divorce probabilities can be calculated by length of the marriage. In France, the probability of divorce is greatest at around five years of marriage (2.9% in 2010), after which it gradually decreases (under 5 per thousand beyond 34 years of marriage) (Mazuy et al., 2011).
The presence of children, the age at marriage and the employment status of the spouse may have a positive effect on the occurrence of divorce risk (Kalmijn and Poortman, 2006).
Divorce risk is not necessarily higher among the less advantaged social categories. For example, women with higher-level or intermediate occupations have a higher risk of separating than female clerical and sales workers. In contrast, men with higher-level or intermediate occupations have a lower risk of separating than male clerical and sales workers (Vanderschelden, 2006).
For example: family benefits, family support allowance (ASF), back-to-school allowance (ARS), lower taxation for lone parents.
In some of these countries, such as Canada, this rule also applies to couples who were not married.
The joint taxation system for couples reinforces occupational inequalities between men and women. When the couple is on an equal standing, i.e. earning the same income, the joint taxation regime is no more advantageous than an individualized one. However, the greater the inequality between the partners, the greater the tax advantage. Consequently, today’s joint taxation regime does not encourage the partners to equalize their incomes.
In some countries, such as Canada, the state guides the choice of the judge by drawing up guidelines with criteria that must be respected when calculating the amount of alimony.
In the USA, some insurance companies propose policies that guarantee continuity of payment of child support or alimony in the event of death or disability of the payer.
Bianchi et al. (2014) show that in the event of a divorce in Italy, the parent with custody of the children (i.e. the mother in most cases) has the right to remain in the family home, even if the other parent continues to pay the mortgage on that home.
According to Jenkins (2008), the narrowing of the gender gap in post-divorce standard of living observed in the UK in the early 2000s was caused in part by an increase in women’s employment rate.