1At the intersection of demographic, cultural, welfare state changes, Italian women now entering old age are facing unforeseen demands and tensions. More of them are still in the labor force than in previous generations, due to the fact that women now in their fifties and sixties were protagonist of the first increase in women’s labor force participation since the post-war years. Furthermore, due to changes in the pension system and retirement age that have particularly affected female workers, they also have to remain longer in paid work than the previous generations of working women. At the same time, given the scarcity of care services, they are under the dual pressure of care demands from above (their own parents or parents-in-law, their spouse) and from below. They do have fewer grandchildren than their own mothers and grandmothers, given the long-standing fertility decline. But these few grandchildren more often than before have a working mother who needs some help to deal with care and supervision needs.
2The present article addresses these questions from a socio-demographic perspective. After reviewing the international literature on intergenerational caring, I will describe care policies in Italy within a comparative context. Based on available statistical data, I will then analyze the changing demographic configuration of intergenerational care across different cohorts of women and the impact of recent pension reforms on this configuration. In particular, I will discuss the emerging sources of tension in intergenerational care relationships that are still largely premised on the availability of middle-aged and young-old women, i.e., those in the 50-65 year-age bracket.
Intergenerational care relationships in the EU
3Sociological and policy research on both grandparental care and care for frail older parents or parents-in-law reveals that both types of care are both very widespread and highly gendered throughout Europe. It also shows that both care relationships exhibit two opposing North-South gradients in terms of frequency and intensity. The frequency of elder care is greater in Northern Europe, while the intensity is more elevated in Southern Europe.
4Grandparental care plays an important, widespread role in contemporary Europe [Di Gessa et al., 2015]. Some studies [Igel and Szydlik, 2011; Hank and Buber, 2009; Bordone et al., 2017] have found that the probability of grandparental childcare of grandchildren under 16 is generally higher in Denmark, Sweden, the Netherlands, and France (approximately 60 %) than in Southern Europe (less than 50 %). Grandparents in Southern Europe, however, provide childcare more regularly (i.e., almost weekly or even more frequently). Grandmothers, particularly maternal grandmothers, are more likely to provide care than grandfathers. Younger and healthier grandparents are more likely to look after their grandchildren [Glaser et al., 2013], particularly if they are not working [Hank and Buber, 2009; Igel and Szydlik, 2011].
5Parental characteristics, family size, and the age of grandchildren are also factors in determining the level of grandparental childcare. The probability of grandparental childcare is highest if the mother is employed [Bordone et al., 2017]. It is also more frequent with grandchildren aged four to six. The most intensive care is provided to children under the age of three [Igel and Szydlik, 2011; Keck and Saraceno, 2008].
6Due to the interaction between low fertility and increasing life expectancy, while the number of small children within a family network requiring care is diminishing, the number of family members in potential need of long-term care due to frailty in old age rose by 30 % between 2013 and 2016 according to Eurostat estimates (EC 2015, 2016).
7Most care of older people is still provided informally and for free, usually by partners and children [oecd, 2011; European Commission, 2016]. As noted in the most recent European Commission Report on Health and Long-term Care (LTC) [EC, 2016: 190-191], informal care represents an “informal economic sector” in EU member-states that is estimated to represent the equivalent of 50 % to 90 % of overall costs of formal ltc. Informal care is thus a highly significant component of ltc that contributes to its sustainability in terms of the public budget, while also posing sustainability issues at the levels of the household and kinship.
8Informal long-term care is also strongly gendered, with women—wives, daughters, and mothers (in the case of disabled children)—performing most care, particularly when it is intensive [Schmid et al., 2012]. Women also ensure the management of care when homecare services or paid personal care are used [Rosenthal et al., 2007]. Men tend to devote less time than women to caregiving or to provide specific, gendered types of help [Gerstel and Gallagher, 2001]. Since men’s informal care predominantly involves caring for older spouses [Bracke et al., 2008], the share of caregiving men increases with age, in a phase when people are typically no longer part of the workforce [Dahlberg et al., 2007; Kahn et al., 2011]. The elder care gender gap within families appears to be consistent across countries [e.g., Jegermalm, 2006; Hanaoka and Norton, 2008; Toffanin, 2011; Da Roit, Hoogenboom and Weicht, 2015]. In contrast with the situation regarding motherhood, however, the existing literature on mid-life (40-60 year-old) women with frail elderly parents reveals that caregiving does not have a significant impact on their employment [Da Roit and Naldini, 2010]. Indeed, any effect on caregivers’ employment tends to involve reduced working hours rather than complete withdrawal from the workforce [Pavalko and Artis, 1997; Spiess and Schneider, 2003]. The most negative impact of providing care for family members has been noted among families that lack adequate financial resources to cope with long-term care needs [Sarasa and Billingsley, 2008], those for whom support is unavailable or unaffordable [Lechner and Neal, 1999; Saraceno, 2010], those who are providing care for a co-resident dependent relative [Heitmueller and Michaud, 2006], and cases involving particularly intensive needs [Crespo, 2006].
9Informal elder care has been found to have a greater negative impact on women’s employment opportunities and working hours in Southern European countries than in Northern countries, with Central European countries located in between the two extremes [Saraceno, 2010; Kotsadam, 2011; Naldini et al., 2015]. The difference is largely due to differences in the incidence of intensive informal care, which is comparatively low in Northern Europe and higher in southern countries, with Central Europe again falling in between [Crespo, 2006; Saraceno, 2010; Kotsadam, 2011; Naldini et al., 2015]. Lower negative impacts on women caregivers’ employment in northern welfare states can be explained by such macro-level factors as higher rates of employment among older working-age women, higher quality and more widely available formal care, and less rigid gendered-care norms.
10Da Roit et al. , specifically examined between-country differences in the family care gender gap. The study concluded that the most important explanatory factor was the institutional framework, when comparative cultural orientations and labor market conditions are taken into account. Based on the analysis of Share data, in fact, the authors found that the absence of adequate social services was the factor that best explained the gender informal care gap [see also Schmid et al., 2012].
11The current preponderance of informal family care for both children and long-term care throughout EU and oecd countries, however, should not be interpreted as a given or projected into the future. Indeed, demographic dynamics suggest that the families’ ability to provide informal care is likely to diminish across the board [Murphy et al., 2006; Haberken and Szydlick, 2010].
12In addition to probable demographic changes, a range of other factors are likely to contribute to a potential reduction in the pool of potential informal family carers. First, given that the vast majority of family carers are women (70 % to 90 % according to 2011 oecd estimates), the pool of potential carers is likely to shrink because of the increased participation by women in the labor force, including longer hours and an extended number of years. Although increasing participation by women in the workforce also increases demand for child care, their participation also reduces their availability as informal carers across kinship ties, whether for grandchildren, elderly parents or in-laws, or other family members with serious disabilities. Second, the spread of unstable partner relationships reduces access to informal support by one of the partners in a relationship and, particularly in the case of men, also by children [e.g., Albertini and Saraceno, 2008]. Finally, those who have never been married/partnered or have no children are particularly vulnerable to lack of candidates for informal family carers [Dykstra and Hagestad, 2007; Albertini and Mencarini, 2014].
13The interplay of these different dynamics and phenomena means that, while the number of children needing some non-maternal care and of family members potentially needing long-term care is expanding, the number of potential family carers is not increasing at a comparable rate. Most European countries address this potential unbalance through various combinations of services (defamilialization) or leave programs and care allowances (supported familialism). These combinations differ not only across countries, but also depending on the type of care, whether it is for children or elderly and disabled individuals. In turn, this contributes to different balances between defamilialization, supported familialism, and familialism by default [Saraceno, 2010; Saraceno and Keck, 2010]. The larger the space left to familialism by default, the greater the pressure on families, particularly on women family members.
The Italian example
14From a comparative perspective, Italian care policies may be described as striking an uneasy balance between familialism by default (or unsupported familialism) and supported familialism, with a small space for defamilialization. Furthermore, both defamilialization and supported familialism are greater in the case of pre-school age children than in the case of the disabled and frail elderly individuals. In the former case, in fact, working mothers are entitled to five months of full paid leave when they bear a child, plus additional 6 months at 30 % of pay, that may be used, also part time, until a child turns 12. An additional 4-5 months are available for fathers. Childcare services for children under three are scarce and unevenly distributed across the various regions. Figure 1 shows how far the various regions are from achieving the EU goal of 33 %. On the contrary, pre-school for children 3-5 is nearly universal, and fees are low. All-day elementary school programs are primarily available in the central northern regions and are not universal.
15In Italy, disabled individuals of any age who are unable to live autonomously are entitled to a non means-tested allowance (indennità di accompagnamento). Homecare services are relatively scarce, however, and generally means-tested based on household income, as well as the (theoretical) availability of family members. Institutional care is also reduced (and expensive) and caters primarily to disabled individuals who have no families or who cannot be cared for in the home. There are also substantial regional differences in both categories of services, with the South most lacking [istat, 2014a]. Figure 2 shows the gap between the Italian and EU contexts in terms of long term care.
Figure 1 – Childcare services gap (for children aged 0-2) between Italian regions and the EU 33 % threshold.
Figure 1 – Childcare services gap (for children aged 0-2) between Italian regions and the EU 33 % threshold.
Figure 2 – Country-specific long-term care recipients, as a percentage of the dependent population. 2009-2013.
Figure 2 – Country-specific long-term care recipients, as a percentage of the dependent population. 2009-2013.
16Workers who care for another family member are entitled to one day off of work per month. Working parents of severely disabled children of all ages and working spouses and children (when there is no available spouse) of severely disabled persons also have access to a six-month leave. This, however, is compensated through the pension system, on condition that the care recipient is the child of the carer.
17The incidence of familialism by default is therefore greater in the field of care for disabled and frail elderly individuals than in the case of small children; it increases with age after the age of three. Furthermore, there are substantial regional differences. Center-north regions on average have higher levels of child-care coverage, more all-day school programs, and more homecare services than those in the southern or in the island regions.
18As long as women’s labor force participation remained comparatively low, heavy reliance on familialism by default was never perceived as particularly problematic, particularly among women with preschool-age children and middle-aged women. Most underage children could count on a “full-time mother.” On the other hand, there was, and to some degree still is, a large pool of “inactive” grandmothers, wives, and daughters to help with the care of grandchildren, disabled adults, and frail elderly persons.
19Grandparents, especially grandmothers, are an important resource for young couples. However, comparative research shows that it is only in Mediterranean countries that grandmothers represent a systematic and often primary work-family conciliation instrument for working mothers [del Boca et al., 2005; Brilli et al., 2013; Keck and Saraceno, 2008]. Grandmothers’ availability also has a positive impact on fertility choices [Aassve et al., 2012]. Istat data [istat, 2014b; see also Mencarini and Solera, 2015] indicate that in 2012-2013, 51.4 % of children from 0 to 2 years old were cared for by grandparents, principally grandmothers, while mothers were wage-earning employees, compared to 37.8 % attending childcare services and 4.2 % being cared for by babysitters. These findings have been confirmed by qualitative research [e.g. Musumeci et al., 2015].
20An istat survey in 2011 [istat, 2012] found that over 55 % of non-institutionalized disabled individuals of any age received help from co-resident or non-co-resident family members, while over 83 % expected to receive help from family members in case of need. Only 2.6 % received professional or paid homecare. Notwithstanding the recourse to primarily female migrant workers as paid carers in what has been defined the “migrant in the family care model” [Bettio et al., 2007; see also Naldini and Saraceno, 2007], when care support is needed, it is usually provided by (female) family members. A cross-EU survey [Eurofound, 2015] found that working-age Italian women represented the second highest percentage (over 15 %) after Hungarian women, providing care for a frail family member. This group is split almost evenly between those participating in the labor market and those who are not formally employed.
21Over the years, istat data  indicate that, notwithstanding the increasing incidence of the older generation in the intergenerational chain, parallel growth in the proportion of working mothers in the younger generation has brought about a gradual shift from supporting the elderly to supporting young families and children within the informal family care budget. This phenomenon suggests a potential conflict between care demands and family loyalties. The often mentioned “sandwich generation”—are required to care for both their own young children and frail elderly relatives— this phenomenon is actually demographically very rare [Dykstra et al., 2016]. It may be comparatively more frequent, in Italy at least, among “young grandparents” (grandmothers) who are forced to manage the care demands of their grandchildren and their own parents.
22Changes in the demographic structure of the population and therefore of kinship networks, coupled with women’s rising participation in the labor force, the very basis for a highly-gendered familialism by default is being undermined.
Changes across three cohorts of Italian women
23The impact of demographic change on kinship networks and on the upward and downward demands of care can be understood by comparing three cohorts of women when they were in their forties: those born in 1940 (now seventy-seven years old), those born in 1960 (the Baby Boomers, now in their late fifties), and those born in 1970, who are now in their late forties [istat, 2014a].
24At forty years of age, the majority of women in all three cohorts had at least one child younger than age 14, although the total number of children varies widely. While the younger cohort had an average of 1.4 children at forty, the middle cohort had 1.7 children, and the older one nearly 2. From the oldest to the youngest cohort, their age at the time of their first childbirth increased—women born in 1970 became mothers at 30, three and five years later respectively than those born in 1960 and in 1940. Furthermore, recent estimates indicate that 20 % of the younger cohort will remain childless, compared to 13 % in the two older cohorts. A higher proportion of the younger generation will thus not have children when they become old, compared to the older generations.
25In addition, the support network in which these three cohorts participate differs in size and composition. At the age of about forty, the two younger cohorts could theoretically count on an average of five persons—husband, siblings/in-laws—to share in the care of children and of frail elderly relatives, compared to nine in the older cohort. The number of grandchildren and grandparents in the kinship network also varies across the three cohorts. For women born in 1970, there are at least three grandparents and four grandchildren. For those born in 1960, there was the same number of grandparents but six grandchildren. Women born in 1940, however, faced the most complex prospect of an average of 10 grandchildren and only one grandmother.
26Although more grandparents may mean more help with younger grandchildren, longer life expectancy has increased the number of years during which elder-care needs are likely to emerge for a shrinking pool of adult children, and particularly daughters. Furthermore, the increased age at first birth in successive cohorts [Mencarini and Solera, 2015] means that women will be steadily older when they become grandmothers, reducing the length of time that they will be available to care for grandchildren. Due to increased life expectancy, women born in 1970 at forty can expect to spend 22 more years with at least one elderly parent. This represents an increase of four years for women born in 1960 and ten more than those born in 1940. Some of these additional years are likely to be marked by some kind of dependency. Recent estimates indicate that the number of years that individuals are likely to survive with poor health is systematically increasing. By 2050, this extended period of some degree of likely dependency is estimated to be 10 years longer than today [Osservatorio Nazionale sulla Salute, 2017].
27Cross-cohort changes also occurred in labor force participation. Among women born in 1970, 62 % were employed at forty, compared respectively to 50 % and 32 % of those born in 1960 and in 1940. The younger cohort of women, therefore, has been able to rely on more grandparents/grandmothers with greater availability of time when children were young than the older ones, but will have less time to care for a higher number of elderly parents/in-laws. These women will also have fewer children than the older cohorts to count upon, when they themselves will become old and frail.
28In addition to cross-cohort differences, there are significant cross-regional and cross-educational differences. Less educated women and women in the South have lower activity levels and employment rates across all cohorts [istat, 2015].
Changing the rules of the game: the 2011 pension reform
29If we examine the age at which women are likely both to become grandmothers and to have still some surviving parents/in-laws, the employment rate of women in the 50-64 year-age bracket increased by over 14 percentage points between 2004 and 2014. This increase, at over 20 %, was particularly significant in the 55-59 year-age bracket, creating a cohort of women with an overall employment rate of 48 % in 2014. Although women of all ages in Italy still have far lower employment rates than in many developed countries, employment is becoming a normal feature for nearly half of the women over fifty years old. These are the cohorts who closed the gender education gap and initiated the increase in women’s labor market participation during the 1990s. These women have also been the most affected by changes in women’s pension age, which strongly contributed to the increased overall proportion of older workers in the working population, particularly in the case of women [istat, 2015].
30The 2011 Italian pension reform, which was approved under the threat of a financial crisis, significantly restricted age and seniority requirements for retirement, particularly for women in private sector occupations who still enjoyed far more favorable treatment than men and women in public administration. While in the public sector, the minimum retirement for women had already started to be equal to men’s in 2010, following a discrimination ruling in 2008 by the European Court, in the private sector, at 57, compared to 65 for men, it was still lower. Further, the seniority requirement for women retirees was lower. More specifically, before the 2011 pension reform, women could retire if they fulfilled the following conditions: 40 years plus 1 month of seniority (pure seniority pension) and a minimum effective seniority of 35 years (that is, excluding notional contributions for sick-leave and spells of unemployment); or 20 years of effective seniority and a minimum age of 60 (compared to 65 for men and for all workers in the public sector); or a sum of age plus seniority greater or equal to 96, with a minimum effective seniority of 35 years and minimum age of 60. Finally, they could retire at 57 if they had an effective seniority of 35 years and opted for a pension benefit calculated according to the defined contribution formula (introduced in the 1995 pension reform, which gradually phased out the previous, more advantageous, defined benefit system), instead of a combination of the two systems, thus with a lower pension.  This last option was only available to women who would retire by the end of 2015 (the so called “opzione donna,” women’s option). It was intended not so much to smooth out the increase in the pension age, which had already begun since the 1995 reform, as to meet the family-care demands confronted by many older women workers.
31With the 2011 reform, the phasing in of the defined contribution system was accelerated and the requirements for retirement were further strengthened, more so in the case of women. The seniority requirement was raised to 41 years plus 1 month and a minimum effective seniority of 35 years, or a minimum age of 62 with 20 years of contribution. The “opzione donna” remained, limited to those who retired within 2015. Age and seniority requirements will gradually increase to align by 2018 with men, whose age is also revised every two years to align to that of life expectancy.
32The debate concerning a lower retirement age for women predated the European Court ruling. Those who defended it, including trade unions and many feminists, argued that it was a compensation for dual (paid and unpaid) work, although the tradeoff involved a lower pension. Women (including this author) who supported equalizing ages argued that, first, not all women had dual working lives, and second, instead of asking for a lower pension age, women should demand more services and better work-family reconciliation policies. Indeed, when women’s retirement age was equalized with men in the public sector in 2010, it was stipulated that the savings produced by the change be invested in childcare and long-term care services. This did not happen, however, because, under the pressure of the financial crisis, the Berlusconi government incorporated those savings into the public budget while reducing the funds for services.
33The complications for work-(extended) family strategies of many female workers caused by the 2011 restrictions on age and seniority requirements are documented in a study on post-2011 absenteeism patterns among older female workers affected by the reform [Coda Moscarola et al., 2015].  The researchers point out that, according to the literature on absenteeism, perceived high strain at work and low social support are effective predictors of sick leave [Andreassen and Kornstad, 2010; Moreau et al., 2004]. It has also been shown that the costs associated with being absent from work significantly affect work absence behavior [Johannson and Palme, 1996; 2002].
34With specific regard to Italy, Scoppa  and Scoppa and Vuri  have noted higher rates of sick leave among workers with higher seniority and more stable contracts, employed in public sector or in big private firms, and living in regions with low unemployment levels. While Scoppa and Vuri explain their findings as the result of opportunism in a country with low controls and high employment protection, Coda Moscarola et al.  focus on work-family conciliation issues. Administrative data do offer evidence of higher sick-leave absences for women obliged to postpone retirement by the 2011 Italian reform and who, in 2011 (before the reform), had already experienced a sick leave period, with a significant direct correlation between weeks of absence and years of retirement delay. At the same time, women who did not already have a sick-leave spell in 2011 reacted to the postponement of retirement only if they were grandmothers. The first finding cannot be unequivocally interpreted. It may suggest these workers’ deteriorating health but also the presence of care demands within their households or extended families. The second finding, however, clearly suggests that “playing sick” is a common strategy for coping with grandparental caregiving demands.
35Although they do express overall support for the 2011 reforms, the researchers conclude that, “notwithstanding that pension reform was needed to recover the financial sustainability of the pension system, it certainly had stringent effects on many Italian workers not too far from retirement, and on women in particular. We cannot say… that our findings concerning grandmothers point to opportunistic behavior. Indeed, a careful consideration of our results seems to support a different thesis. Italy suffers from a chronic lack of well-structured high-quality care facilities, and middle-aged women are often called to stand in. Sick-leave may then be the response of last resort.”
36There is another, more qualitative, indicator of the imbalance caused by the 2011 reform on work-family conciliation strategies of older women workers in a policy context like the Italian one, where, due to austerity measures, already insufficient services were increasingly underfunded, reduced, or made more costly for users. Many women, supported by trade unions and some bipartisan politicians, lobbied the government in favor of continuing the opzione donna beyond 2015, even if it implied, on average, a loss of about 30% of their pension incomes. This request was not accepted. The 2017 fiscal policy, however, experimentally introduced the possibility of early retirement for workers with a minimum of 30 years of contributions, who were 63 years of age, and who had a maximum of three years and seven months to go before reaching the official retirement threshold and could count on a future pension at least one and a half higher than the minimum.
37The pension benefit during this “anticipated” period would be financed through an insured bank loan to be repaid by the pensioner through a reduction in the pension benefit once he or she reaches the full pension age. In this complex (and costly, for the interested pensioners) system, family caregivers (spouses or parents) who live with cared recipients have been included among those—the long-term unemployed or those employed in highly risky occupations for at least six years—allowed to retire early without having to repay the loan later on. In this case, however, the benefit cannot exceed 1,500 euros per month, regardless of the level of the future pension. The carer must also be able to demonstrate that she or he has had this responsibility for at least six months before applying for the benefit. Caring “across” households, either for the frail elderly relatives or grandchildren, is not accompanied by any entitlement, and the only option is a very costly bank loan scheme.
38Women in Italy who are now in their fifties or early sixties are the forerunners of a slow but steady process of change at the demographic, family, and gendered behavior levels. They have entered the labor force, and continued to work while having a family in greater numbers than earlier cohorts. At the same time, their kinship networks are experiencing extended intergenerational ties and growing demands for care among frail elderly relatives or in-laws and grandchildren whose mothers remain in the labor force. Policy arrangements in the area of care, however, appear unable to keep pace with these changes and largely premised on an idealized stable family unit that is defined by a strong gender division of labor and a substantial amount of available uncompensated caretaking time. Ironically, recent changes in labor market regulations and the pension system undermine women’s availability to care for relatives. In particular, the recent pension reform, which has increased the pension age of women at a time when care services are increasingly limited by austerity measures, further eroded the availability of (uncompensated) women’s time that was already reduced because of demographic changes and women’s increasing participation in the labor force.
39In the mid- and long-term, the combination of demographic changes, women’s participation in the labor force, rising pension ages, and sub-minimal defamilialization could very well provoke a societal “care deficit,” particularly with regard to the elderly, who have fewer available services than young children. As a consequence, social inequalities could increase between those who live in better served regions and those who live in poorly served areas, and between those who have enough resources to turn to the private care market and those who do not.
40To counter this risk, the familialistic approach to care policies should be corrected, particularly with regard to long-term care. Although it is true that the need to partially defamilialize childcare has marginally entered the policy agenda through both work-family reconciliation and social investment in children’s services, long-term care remains on the margins of policy discourses and continues to reflect a static vision of the family, of its demography and gender-based division of labor. Even the “new,” limited policies that do address the pressure of family care demands on older women workers for family care seem inspired by a persistent, familialistic, strongly gendered conception of who should provide care. Because financial constraints do not allow these policies to be expanded, problems experienced by older female workers in managing their caring responsibilities may justify less familialized policies, not only in terms of child care, but also long-term care by shifting the focus to providing more services. ■
The 1995 reform stipulated a gradual passage from a defined benefit to a defined contribution period. In effect, it divided workers into three different groups depending on seniority as of December 31, 1995:
– Defined Benefit (DB) workers, i.e., workers with more than 18 years of seniority, were entitled to maintain the rather generous DB formula, also for future seniority;
– Pro-rata Defined Contribution (pro-rata DC) workers, i.e., workers with less than 18 years of seniority whose pension benefits would be calculated according to a pro-rata mechanism (DB for past seniority and DC for future seniority);
– DC workers, i.e. new entry-level workers whose pension benefits would be computed entirely using the DC formula.
All older workers in 2011 were in either the first or in the second group.
It should be noted that in 2011 one of the researchers, Elsa Fornero, was the Minister of Labor in the Cabinet that introduced this reform and the person identified as its principal author (the reform is called “Fornero reform”).