Cette page existe uniquement en anglais

Home > Activities > Committees >Age Structure and Public Policy > Meeting Report

International Meeting on Age Structure Transitions and Policy Dynamics:
The Allocation of Public and Private Resources Across Generations

Taipei, 6-8 December 2001

Organized by the IUSSP research committee on ’Age structure and Public policy’ and the Institute of Economics of Academia Sinica.

Report

The third conference organized by the IUSSP Committee on "Age Structure and Public Policy" and Academia Sinica took place in Taipei, Taiwan, in December 2001. The purpose of this third meeting was to examine issues related with age structural transition in relation to policy dynamics. The conference included seven sessions, which looked at different aspects of intergenerational transfers. The topics covered in these sessions included: the allocation of resources to children and the elderly; factors associated with intergenerational transfers, such as fertility decisions and other demographic events; Public spending in education and health; patterns of giving across generations in developed and developing countries; financial and demographic macro-level perspectives; and policy implications.

1. Child and the Elderly: The Public and Private Dimensions

The three papers presented in this session looked at the allocation of resources to dependents. The first one focused on changes in the provision of Public transfers to the elderly in relation to children. The other two papers looked at the set of mechanisms (social contracts and motives) affecting the allocation of private transfers.

In the first paper, "Social Expenditures on Children and the Elderly, 1980-1995:Shifting Allocations, Changing Needs", Janet Gornick examined possible shifts in the allocation of welfare state expenditures between 1980 and 1995. She addressed two weaknesses that from her point of view pervade the welfare state literature. First, changes in welfare state provision have not been adequately disaggregated. Second, most assessments on changes in welfare state efforts have failed to take in account changes in levels of need. using information from 14 industrialized countries¹, the author compared the expenditures on the elderly relative to children and evaluated changes in needs (changes in population shares and in market earnings). The period under study was marked by considerable changes in demographic and economic conditions. The elderly-child ratio increased in all countries (in some countries more than in others) as a result of a sharp increase (24 percent on average) in the number of elderly persons and a decrease in the absolute number of children. At the same time, the GDP growth outpaced population growth, so GDP per capita grew steadily in all countries. The share of social spending allocated to children benefits compared to elderly benefits changed in diverse ways across countries, but remained unchanged as a share of total social expenditures (28 percent and 7 percent, respectively). When controlling for shifting population shares, the author found that both type of expenditures increase quite sharply in ten of the 14 countries. The outlays for the elderly grew on average 43 percent (except in Australia), and the amount spent on families with children rose 87 percent (except in Germany, Italy, and the Netherlands, where cash benefits per child even decreased). Finally, changes in social expenditure relative to market income -as an indicator of need- increased modestly (on average 5 to 6 percent). However, when the change in social transfers relative to this indicator of need was disaggregated, the author found a larger expansion for families with children (about 19 percent) compared with those for the elderly (about 2 percent). In sum, this study did not find evidence of welfare state reversals between 1980 and 1995. The author suggested that studies on welfare state should focus more on changes experienced by different population groups and should extend the traditional measures on welfare state effort.

Nancy Folbre’s theoretical paper, "Social Family Contracts and Intergenerational Sharing", offers a framework for understanding some aspects of the distribution of the costs of dependents. She argued that allocation of resources is shaped by social family contracts, which include sharing rules across and between generations. These contractual arrangements do not emerge directly from individual optimization processes, but rather reflect an efficient system of social reproduction and the relative bargaining power of groups. She developed a model of intergenerational transfers to evaluate family contracts with respect to efficiency and fairness. In this model, workers agree to share their output with dependents (elderly and/or children). If the percentage shared remains constant through generations then a simple form of intergenerational reciprocity is guaranteed (fairness). However, many factors alter these patterns of intergenerational sharing (technological change, change in the relative size of generations, or changes in bargaining power). The author called for more attention on the ways in which an implicit social family contract affects the distribution of output among generations. This paper suggests that social engineering, i.e. collective action that establishes rules, norms, and other social institutions, significantly affects the distribution of the costs of caring for dependents.

In the discussion of the first paper, Anne Gauthier suggested including poverty rate in the analysis to assess if the trend on poverty rate noted in Preston’s 1984 paper has occurred in the countries under study. That is, it would be interesting to evaluate whether the shift of Public resources from children to the elderly has increased the poverty rate of children and decreased that of the elderly. Additionally, she recommended focusing first in the average of all countries to have an overall picture of the situation and then go on detail to look at each country. She also suggested looking at changes in the composition of the population in further analyses. The participants raised the issue of including services in the analysis to support this thesis. Ian Pool highlighted the importance of taking into account the labor market group, as it might have an important impact on social spending. Regarding the second paper, the discussant wandered if the model presented was a universal proposal. She suggested including more applications to the theoretical model. Ian Pool welcomed the argument that social engineering takes place in the allocation of resources.

The third paper of this session, "Intergenerational Transfers in the Family: What Motives for Giving?", studied the motives of transfer giving. Martin Kohli and Herald Kunemund argued that it is important to assess not only the amount of the transfers, but also the way the monetary transfers are given. Motives are critical to understand the quality of transfers and how people react to changes in transfers. The authors used information from the German Survey of Ageing (retired persons between 40 and 85 years) to assess the structure of transfer motives. Findings from principle components analysis showed that the motivation for giving monetary transfers does not come as a single motive, but instead as a combination of motives that overlap and interact among each other. The results showed that three components or dimensions could explain motivation. The first one denotes unconditional giving, the second one comprises motives that represent giving under certain conditions, and the third (less important dimension) captures independence and separation of the older generation. Thus, the structure of motivation can be explained by a dichotomy of unconditional-conditional giving which does not correspond to the commonly assumed dichotomy of altruism-exchange. Among the predictors of giving, the author found a strong effect of social status and gender. unconditional giving is more frequent among women and among those with higher income. In short, the authors find that motives have a strong impact on the incidence and the explanation of transfer giving.

The discussant raised issues regarding the quality of data. She suggested including macro level data variables in the analysis (size of the community, indicator of regions, type of activity) as these may have an important effect on the results. Some participants also questioned the quality of data. They discussed that there might be a possible bias in the responses: a product of a social desirability effect. The suggestions for further analyses included: comparing these results with other countries, disaggregating results by people giving and not giving transfers, and looking at factors such as social status, social pressure, and preferences.

2. Intergenerational Transfers, Fertility Decisions and the Role of Time

The three papers presented in this section cover different aspects related to the provision of family support. Cyrus Chu and Ruoh-Rong Yu’s paper, "Bequeathable Assets, Social Networks and Family Visits" explained that data from the Taiwanese Panel Study of Family Dynamics shows that the proportion of parents who have divided all their assets is quite high. Furthermore, children’s frequency of visits is higher for those families whose assets have been divided than for those whose assets remain intact or who do not possess any assets. The authors argued that neither the hypothesis of accidental bequests nor that of altruistic bequest could explain this phenomenon in Taiwan. Thus, they propose a social network theory to explain the Taiwanese parent-child interaction. They support the hypothesis that the main driving force of children’s behavior is the norm instead of the assets held by parents. Their empirical results show that children’s frequency of visits is higher when tighter kinship networks prevail. Additionally, the impact of kinship network is greater when parents have divided all the assets. In other words, the social norm is seriously triggered after asset division. Even when controlling for the possibility that parents may know that asset-division might influence children’s behavior, the interaction between kinship network and asset-division still had a significant effect on filial attention. Their findings confirm the importance kinship networks have in explaining the feedback of children to their parents.

The second paper of this session, "Determinants of Two Types of uxorilocal Marriage in Lueyang, China", aimed at understanding the changing family systems and the marriage customs in rural China. Marcus Feldman and his coauthors pointed out that the location of residence after marriage reflects the traditional patrilineal family system in rural China. Virilocal marriage, where the wife leaves her natal family and moves to her husband’s family, occurs almost universally in rural China, while uxorilocal marriage, where the husband leaves his family and moves to his wife’s, has been rare. Although cultural factors have an important role in determining the type of marriage, there are economic and demographic factors that affect the decisions on the type of marriage. In Lueyang, one of the rare places in China where the sex ratio at birth is biologically normal (105 compared with 116 on average in the whole China), the proportion of uxorilocal marriages is relatively high (over 30 percent). Thus, information from this county allowed the authors to study in-depth this type of marriage and to compare the determinants between the two types of uxorilocal marriages: contingent, driven mainly by demographic factors; and institutional, driven by practical economic purposes. Findings from this research show that uxorilocal marriage is more likely to occur in families without a son or in those who are unable to adopt a son. The latter suggests that the contingent type is more frequent than the institutional one. The determinants common to both types of uxorilocal marriages are household structure, family clan, and attitudes towards this type of marriage. The only difference is that the effects of the determinants are stronger for the contingent type. Among the unique determinants of the institutional uxorilocal marriage, the most important one is the influence of parental marriage type within the household. Additionally, the authors found that only institutional uxorilocal marriage is transmissible within a household and that transmission has been affected by the 1978 rural economic reform. The low fertility and the increasing number of no-sons families in China will continue to influence marriage customs. The authors conjecture that uxorilocal contingent marriage will increase across rural China in order to provide family support for the elderly who are not covered by any social security system.

In the discussion of the first paper of this session, Robert Schoen remarked that the two-stage model used in this analysis shows the power of utility models in illustrating social forces. He asked about the possibility of including other important elements in the analysis such as number of assets, sibling interactions, favorite sons, and co-residents. Other participants were interested in further analyses on changes over time. The author was asked whether family obligation is the only norm that rules the asset division behavior in Taiwan. A final comment suggested that for issues on transfers the researcher should consider the emotional aspect of decisions and not only the rationale choice model.

Regarding the second paper, the discussant commented that results of this study show that marriage customs in China are a mechanism for preserving the family system. One of the participants asked which institutional mechanisms affect these decisions. It was suggested to disaggregate the results according to socio economic differences.

At the end of this session, Sumon Kumar presented the paper "Intergenerational Transfers: the Ignored Role of Time". He indicated that literature on intervivos and intergenerational transfers has focused on the possible determinants of transfers but that it has overlooked two important aspects. First, it has not taken into account the possibility that transfers may be driven by specific events in the recipient’s lives (marriage, childbirth, illness). Second, it has overlooked the possibility that within a reasonably long time, transfers may influence household’s income as much as household income influences transfers. using information from the German Socio-Economic Panel for 1996 and 1997, the author found that demographic events determine transfers to a significant extent. Findings from this analysis showed that current household income unambiguously has a negative impact on receiving private transfers, while the demographic events unambiguously have a positive effect on such transfers. Events like marriage, divorce, and childbirth increase both the probability and quantum of private transfers. This fact suggests that savings are aimed not only at supporting old age consumption but also at ensuring children and other members of dynastic families against negative income shocks and positive expenditure shocks. Thus, overlooking the endogeneity of income may lead to bias in the econometric estimations.

The discussant stressed the importance of including time in this type of analysis. He added that it would be wise to look at lifetime of giving to have a better picture of what goes on in the family over time. Questions were raised on the impact of demographic events on transfers. The participants wanted to know if these events are exogenous shots or anticipated events that trigger transfers.

3. Education, Heath and Intergenerational Transfers

The first two papers presented on this third session analyzed the impact of Public spending on health and education. The third one focused on the changes in private transfers as well as the factors associated with this pattern of transfer giving. Cem Mete and Paul Schultz’s paper, "Health and Labor Force Participation of the Elderly in Taiwan", examined Public health spending and the efficiency of various schemes for financing Public and private health care. The authors assessed the impact of health status on the labor force participation of the elderly and whether the national expansion in health insurance in 1995 in Taiwan encouraged early retirement and hence lower labor force participation. They explained that provision of universal health coverage has been largely debated, as some argue it might have positive effects and others argue the contrary. A national health policy as such may contribute to a healthier population through the use of more health care. Conversely, it may reduce labor force participation and thereby have a negative effect on national income. using data from the 1989, 1993 and 1996 Surveys of Health and Living Status of Middle Aged and Elderly in Taiwan they analyzed the differences before and after the implementation of the National Health Insurance (NHI). One of the main changes observed was a reduction on the share of household total expenditures spent on health goods and services. This decline was even greater among lower income households, which suggests that the NHI scheme achieved a more equal distribution of health care in Taiwan. Regarding health status, the authors did not find any consistent evidence of improvement or deterioration between 1989 and 1996. Additionally, the analysis did not show a statistically significant reduction in labor supply after the establishment of the NHI program in Taiwan. The response among those most likely to benefit from NHI was to increase labor force participation for women and decrease participation for men, although estimates were not statistically significant. This study demonstrates that universal heath coverage did not have a negative impact in the labor force participation of the elderly or in the national economy.

Mark Hayward argued in his paper, "Consequences of Educational Changes for the Burden of Chronic Health Problems in the Population", that as fertility and mortality levels continue to decrease, many developing nations will face the dilemma of investing in benefits for an increase old age population in opposition with investing in benefits for children. He sustained that investments in children’s well-being pay substantial dividends decades later when children become the elders of population, both in terms of reducing the burden of elders’ health problems at the societal level and in improving the quality of life for individuals. He emphasized the growing recognition that education has far-reaching multiplier effects and that some of these effects are manifested decades later in the reduction of major chronic diseases such as cardiovascular diseases or diabetes. His analysis showed that populations that undergo the education transition from primary schooling or less to the completion of high school are likely to experience a substantial decline in the burden of disease. Furthermore many years of life are added to the average life cycle and the vast majority of these years appear to be years without major functional limitations. In sum, individuals gain longer, healthier lives while the collective costs of health care to future generations of elderly are reduced. He concluded his paper underlying Preston’s argument that society as a whole gains more from a life course perspective than from a generational perspective.

In the discussion of the paper by Cem Mete and Paul Schultz, Shripad Tuljapurkar praised finding evidence that the NHI in Taiwan did not decrease labor force participation. He asked about the effect of social security over time and if it was possible to understand the secular changes over labor force participation with this data. His comments for the second paper welcome the explanation on the relation between health, education, and mortality. The questions referred to the impact of education throughout time, how changes in education affect this relationship, and the comparison throughout time and between different places. Mortality, education and health status are correlated products of economic and cultural settings. Some participants questioned the association between education and health. They argued that education might be part of a stratification process. Moreover they questioned the effects of education transition in health in an egalitarian society. Another issue raised was the need to control for family background in order to avoid overestimating the effect of other variables. Lastly, participants asked about the influence of cohort effects on the results.

The study by I-Fen Lin and coauthors provides a dynamic view of transfer giving in their paper "Stability and Change in the Pattern of Intergenerational Transfers in Taiwan". They used two waves (1989 and 1999) of a unique dataset, the Survey of Health and Living Status of the Elderly Taiwan, to assess stability and change in intergenerational transfers. The authors found that approximately two-thirds of the elderly have a stable pattern of financial transfers, with 47.6 percent consistently receiving support and 16.8 percent consistently receiving no support at both survey dates. Another finding relative to changes over time is that a higher percentage of elderly loose support than gain support. A second purpose of the study was to determine which demographic, social, and economic characteristics of the older parents were associated with the pattern of transfers. The results showed that children’s provision of financial support largely depends on the needs of their parents. Parents without higher education, those who are not working, those without pensions or assets are more likely than their counterparts to receive financial assistance from their children. The larger the number of children, the more likely the parents will receive support from them and the less likely the parents will lose support over time. Parent’s marital disruption reduces the likelihood that they will obtain financial support from their children over time, as compared with married parents. On the other hand, changes in elderly parent’s needs (death of a spouse, change in health status, change in assets, among others) are not related to change in parent’s receipt of monetary support. Parents who continue to work are less likely to obtain children’s support, as compared with parents who are not working at baseline or follow-up. Moving from non-residence to co-residence is associated with gaining support. Parent’s acquisition of pensions is associated with losing financial support from children.

The discussion of this paper by Diane Macunovich stressed the need of reporting results as descriptive and not as causal because of the type of analysis done. She indicated that more controls in the analysis were needed. First, controlling for other factors that may explain certain causal relationships. Second, controlling for attrition bias due to mortality as results may belong only to the healthy population. She suggested looking at other type of supports other than financial. Participants proposed including other type of controls: presence of grandchildren and type of location (rural or urban area). The last issue raised was timing, i.e. the duration of the support. This is an important question for policy matters as it is important to know for how long the elderly keep their supports.

4. Intergenerational Support and Transfers in Developed and Developing Countries

Regarding this topic, the first two presentations examined models of intergenerational support in two developing countries: Malawi and Indonesia. The third presentation compared the exchange patterns of 20 industrialized countries. Alexander Weinreb presented a study looking at the importance of lateral components on intergenerational support structures (relations among uncles and nieces, aunts and nephews). In "Substitution, Substitutability and Complementarity: The Effects of Kin Availability on Intergenerational Transfers in Malawi", he argued that in certain societies individuals might have more intensive structured ties with second-order blood relations than with first-order blood relations. Overlooking lateral relations implicitly privileges biology over culture and also ignores a key premise of multiple approaches to the analysis of exchange. using data of Malawi this paper explored how individuals choose to shape their intergenerational support network out of a universe of people normatively defined as kin. The results obtained imply that individual’s transfer relationships are at least in part contingent on the overall structure of the kin network. Kin availability seems to affect the likelihood of transfers much more than it does their value. This is useful to identifying strategies -including transfer strategies- that individuals use in response to demographic changes with respect to kin networks. The author explained the importance of this type of study at the micro and macro level. It allows to understand how individuals tend to structure their exchange networks in response to particular demographic constraints such as lack of father, maternal uncles, and so on. And, it enables to understand how these aggregated individual actions appear to signal emergent structures of jural obligations that are partly the product of demographic constraints.

Elisabeth Schroder and Philip Kreager’s paper, "Age-Structural Dynamics and Local Models of Population Ageing in Indonesia", explored networks over time in three villages in Indonesia in order to uncover patterns between networks and social constructs. She noted that Indonesia is a case of interest for aging, as it has the tenth largest elderly population in the world and the number of elderly persons is likely to increase fourfold over the next three decades. Cohort imbalances persist over most of the 20th century, and strongly suggest that current levels of population aging are due to a significantly different set of demographic factors that will shape the course of age structural transition 20 years from now. Infertility, divorce and migration are among the main disturbing factors that have shaped the size of currently elderly cohorts and the support networks available to them. The authors indicated that the macro-demographic picture provides only the outer contours of current and future intergenerational relations. This paper aimed to provide a more detail picture of the daily problems faced by the elderly. The comparative study showed the following results: that supports tend to flow from older to younger generations, even when elders are quite old; that co-residence is likely to reflect insufficient income in the younger rather than the older generation; that economic strata are not in themselves a reliable guide to elderly options and behavior; that there is no single reliable rationale for describing which children will assist poorer parents; and that the open duration and timing of support amongst members of a kindred are features which require much more attention than prevailing survey and modeling of economic-demographic interactions generally give.

The discussant, I-Fen Lin, stressed the importance of looking at lateral supports as up to now little attention has been given to this type of support. She asked whether the financial transfers analyzed were regular or just due to an emergency. Another issue raised by the discussant was related to the family hierarchy established for giving supports. Regarding the second paper, she found the results obtained very interesting. She suggested incorporating more family background information into the analysis. She questioned how to generalize results from qualitative data. Both papers were asked on the way to include a dynamic perspective of support in their analyses.

"Maternal Co-residence and Contact: Evidence form Crossnational Surveys" by Judith Treas and Philip Cohen focused on the exchanges between grownup children and their mothers. The type of exchanges examined included the likelihood of shared housing and the frequency of face-to-face interaction. This study used data from the International Social Survey Program with information from 20 industrialized countries². The results showed substantial cross-national variation. Not only does the likelihood of living with a mother vary, but also the frequency of visiting a mother with whom one does not live differs from country to country. Another factor associated with intergenerational relationships is gender. While men are more likely than women to live with their mother are, women visit their mothers more frequently (although gender differences are not as marked as differences in co-residence). The private intergenerational exchanges captured by maternal co-residence and contact exist in national contexts with distinctive histories of cultural tradition, religious heritage, and state policy. Balance of Public and private efforts is most evident at the extremes. In the social democratic countries of Norway and Sweden, where Public support is most fully developed, there is less intergenerational contact and markedly less maternal co-residence. By contrast, maternal co-residence and contact is high where de-familiarization by the state is limited (the former socialist countries, and Italy). The author underlined that although there is an association between Public and private transfers, they are not interchangeable.

The discussion by Martin Feldman on Treas and Cohen’s paper noted that the main variable of this study, co-residence, may have problems in interpretation as it is difficult to define or establish its boundaries. The discussant agreed on the importance of understanding the impact these patterns have on fertility as the former will have policy implications.

5. Financial and Demographic Macro-level Perspectives

The four papers presented in this session look at the macro aggregate effect of demographic and financial factors on the patterns of intergenerational support. Juha M. Alho and Reijo Vanne argue in their paper "On Predictive Distributions of Public Net Liabilities" that part of the inheritance left to the future generation is the positive or negative Public net wealth. They explained that to evaluate the burden left to the future generations, intertemporal Public liabilities (IPL = discounted future entitlements- discounted future taxes- current Public wealth) should be viewed as a random variable, and applied the former to the evaluation of the burden we leave to the future generations. That is, IPL calculations should treat all future values as random variables since their results are inherently uncertain. using a stochastic forecast method, they compared the different sources of uncertainty relative to Public liability: future population, future earnings (taxes, pensions), future unemployment (entitlements), future value of current wealth (stocks can have dramatic changes from day to day). The resulting variance corresponding to the share of demographics was 30 percent. The rest was due to economics. The share of uncertainty of future taxes and entitlements was 54 percent, and the remainder 15 percent was due to uncertainty of initial wealth, under the fixed value portfolio policy. Findings from this study demonstrate that current and past generations are not leaving the future generations an inheritance of debt and misery in Finland.

using personal consumption expenditures data in the U.S., Diane J. Macunovich examined the impact aging of the postwar baby boom has had on patterns of consumption and saving in the economy. Findings from her paper "Lessons from the Baby Boom: Delayed Effects of Fertility on Patterns of Consumption" suggest a very complex pattern of age structure effects. The results are consistent with the life cycle model where there is a marked age-related fluctuation in the proportions of income consumed and saved over the life cycle. But at the same time, there is a much wider range of dependency effects. For example, increases in spending due to changes in one age group might be influenced by variations in the size of other age groups. The effect of age structure on personal consumption revealed at higher levels of income a strong u-shaped pattern of consumption expenditures among children under age 15, with strong savings among children aged about five to twelve. The results support the hypothesis that the dependency effect changes with income level: children in lower income economies have an a overall positive (negative) effect on the consumption (saving) rate, but as per capita income rises that effect is reversed. These age-group effects suggest that holding other factors constant, the baby boom generated changes in age structure that have accounted for swings of about 25% in total real aggregate personal consumption demand. The estimates presented suggest that the baby boom has had a major impact on the U.S. economy.

The discussant on Alho and Vanne’s paper agreed on the importance of introducing uncertainty in the models as there is always limited information on the future. He suggested as a useful goal to experiment with different values and criteria when designing a policy. This way it will be easier to decide which is the best policy to implement. On the second paper the discussant suggested incorporating other variables into the analysis such as adult equivalence measures and gender. He noted the importance of looking at the impact of these results in policy issues. Nancy Folbre suggested that both papers could benefit from feminist work. Variables such as women’s labor force participation might contribute to better understand these phenomena. Comments from the participants suggested looking at the economic age structure of the group 15 to 24. One of the participants indicated that model estimation should take into account that this group of age is not only a consumer but also a producer. They suggested the inclusion of two other variables in the analysis: time and regional differences.

In the paper "Age Structure and Aggregate Savings: the Case of Taiwan", Sheng-Cheng Hu and Vei-Lin Chan analyze the magnitude and the factors associated with the observed changes in Taiwan’s household saving system. They argued that since saving behavior is different across ages and generations, the changing age structure has had a severe impact on Taiwan’s household saving patterns. Analysis of cross-sectional household data showed that micro household aggregate saving rate fell by 5 percentage points since 1993. Although the magnitude of this fall is smaller than the decline of macro aggregate saving rate, Taiwan’s households save less. Findings from this study showed that the main determinants of household saving rates are income growth, wealth effect and some demographic variables. The authors found that saving rates have strong negative effects from owning illiquid assets. Furthermore, the younger cohorts, between the ages of 20 and 40, are the ones with higher housing ownership rate. This study confirmed that taking age structure into account is very important for understanding patterns of change. The authors emphasized that to raise household saving rates in Taiwan, faster economic growth is needed.

Robert Schoen and Stefan Jonsson presented the last paper of the conference, "Some Intergenerational Transfer Implications of Birth Fluctuations". using a cyclically stationary population model, they showed that population dependency burdens and individual gains and losses from intergenerational transfers vary substantially with the amplitude of birth fluctuations and the length of each cycle. The author explained that substantial temporal and intergenerational issues could arise in the absence of long-term growth when there are fluctuations in the number of births. At the population level, dependency burdens can vary greatly over the course of a cycle. At the individual level, gains or losses from intergenerational transfers differ dramatically by year of birth. They argued that cohort size is a key factor, with large cohorts profiting under defined benefit systems and small cohorts profiting under defined contribution systems.

The discussant of Sheng-Cheng Hu’s paper, Juha Alho, asked to be more explicit on the definition of savings. He addressed some methodological enquiries and suggested writing down the assumption of the model, specifically those regarding the residuals. Regarding Robert Schoen’s paper, he suggested an alternative analytical approach to estimate and confirm his results.

6. Policy Implications

Ian Pool presented the conclusions of the third conference of the Committee. He reviewed the significance of policy issues, he highlighted the theoretical and methodological issues emerging from the seminar, he gave a detailed analysis of the components of the model emerging from the seminar, and he mentioned the main conclusions from the papers presented in the conference.

He emphasized that policy directed analysis require its own theory and the development of methodologies. Throughout the seminar, Ian Pool stressed the need for a more complex interdisciplinary approach. There is a need to study how to integrate theories of rational choice and efficiency. He noted that a key factor for delivering programs and services to specific age groups is the analysis of age structural transitions. Age structure and policy linkage is relevant not only for the social sector, but also for the financial and fiscal sectors. He underlined that time, duration, period and cohort effects are extremely important for this type of modeling. Regarding the distributional mechanisms of transfers, the papers in the seminar concentrated on the private and Public sector, however there are other mechanisms such as the voluntary sector and the market. He indicated that mechanisms occur at various levels of aggregation (micro and macro level). In the seminar, Folbre indicated that we cannot separate macro from micro effects, and Macunovich noted that macro and micro effects show different dimensions. With respect to theoretical issues, Pool outlined that to date the papers have focused on economic or social theories, but little has been done in trying to integrate both theories. This integration is crucial for a better perspective of the phenomenon under study. Additionally, he underlined that this kind of studies should elaborate issues of cohort timing, period and duration effects for all the determinants, mechanisms and outcomes. On methodological issues, he mentioned the following issues: extending the traditional welfare state measures, taking into account changing levels of need, and disaggregating results by different groups of age. Concerning the data sources, researchers should extend the range of sources used. There should be an improvement in question design and an increase in the number of multilevel, multimethod surveys. The main lessons from the seminar are:

He concluded by saying that this is still a pioneer area, which requires further theoretical and methodological research.

Maria Carmen Huerta

Notes

¹ The countries in the study were divided into three groups: European union countries, European non-Eu countries, and Non-European countries. The Datasets used in this study were the OECD Social Expenditures Database and the Luxembourg Income Study.
² Australia, Austria, Canada, the Czech RePUBLIC, East Germany, West Germany, Great Britain, Hungary, Ireland, Italy, Japan, Netherlands, New Zealand, Northern Ireland, Norway, Poland, Russia, Slovenia, Sweden, and the united States.

Some of the participants of the Seminar