Comparisons of the economic position of single older women in industrialized countries have shown substantial differences, especially regarding the proportion of widows, divorcees, and never-married women experiencing poverty. An examination of these differences may shed some light on the role of different structural features of old-age security plans in limiting poverty. This article focuses on six "first-world" nations that offer a representative variety of system features: France, Germany, the Netherlands, Sweden, Switzerland, and the United States. It makes use of earlier published comparative income and benefit studies, and probes into the economic impact of institutional arrangements of old-age security schemes on aged women who are on their own.
The review comprises four sections:
1. Section I reviews comparative stud-Poverty Among Single Elderly Women Under Different ies that offer information on incomes
Systems of Old-Age Security: A Comparative Review of the elderly, ranking countries in terms of poverty rates for older single
This study takes stock of available comparative research on the economic used will show why consistent com-status of elderly single women in six industrialized countries: France, Ger- parisons are difficult; many, the Netherlands, Sweden, Switzerland, and the United States. A sys- 2. Section II presents findings from tematic comparison of income has become easier due to such standardized comparative benefit studies. Com-data bases as the Luxembourg Income Study. parisons of benefit levels have been
But an explanation for different poverty rates among older women who are included in a few of the research efforts mentioned in the previous
on their own requires a further, differentiated assessment of the countries' section, but mostly, one must turn to retirement benefit structures. This article attempts such a review. It makes different sources when it comes to use of a variety of single-country sources and takes into account the institu- cross-national benefit information. tional heterogeneity of old-age security systems. The study concludes with a Their considerable heterogeneity view of the effectiveness of different old-age security systems in preventing raises additional questions of compa-poverty among older single women. rability, but they are useful for establishing the relative ranking of different systems;
*Dr. Siegenthaler is Professor of Sociology at American University. This article 3. Section III examines the main char-is based on research conducted at the Social Security Administration under the acteristics of old-age security provi-terms of the Interagency Personnel Act. sions separately for each country,
Acknowledgments: The guidance and support of Lillian Liu, Susan Grad, and focusing specifically on measures Martynas YCas were integral to this project. Valuable information and suggestions dedicated to the prevention and alle-were received from Howard Iams, Michael Leonesio, Conception McNeace, Janice viation of poverty among older Olson, Daniel Radner, Denton Vaughan, David Weaver, and Bernard Wixon during women. Here again we aim at comthe course of the inquiry. Translations from the Dutch and Swedish were compe- parability, but must be sensitive to tently provided by Veerle Zanmeel and Landon Gildar. institutional differences that are not always sufficiently appreciated in more general comparative benefit research. The measures assessed include the size of social security benefits, the extent of earnings and
asset income, the role of minimum benefits, the availability of complementary occupational pensions, and the role of social assistance; and
4. Section IV rounds out the comparison and synthesizes the findings. It also assesses the relative effectiveness of different types of old-age security provisions in preventing poverty among single elderly women.
Section I: Comparative Income Studies
Answers to questions about older single women's economic well-being can sometimes be found in general studies of income that use a cross-national perspective. Although conducted with a broader focus, some of these studies are of use for judging the adequacy of social benefits (that is, transfer payments) for older recipients. Comparative income studies are therefore the first, and basic, source reviewed here.
Cross-national studies have frequently paid attention to poverty rates. In some cases, this was due to an explicit focus on poverty. However, comparing relative poverty rates has become a general methodological convention that eases the interpretation of results. We will adhere to that convention.
There are two main lines of relevant research; one based on activities by the Statistical Office of the European Union (Eurostat), the other initiated by the Luxembourg Income Study (LIS). Tables 1 and 2 list comparative studies of economic well-being focusing on single elderly individuals. The order in which they appear represents increasing specificity and completeness of the research efforts for the group being measured.
Statistical Office of the European Union
Early efforts at compiling comparable income data were based on the Statistical Office of the European Union (Eurostat) figures, starting in 1978. Several Eurostat publications have offered comparable poverty statistics on European Union (EU) member countries, often based on household expenditures and referring to the 1980's (for example, Hagenaars et al. 1994). O'Higgins and Jenkins (1990) published comparative poverty rate data based on Eurostat household income statistics. A study of Belgium, the Netherlands, Luxembourg, Ireland, Greece, and the regions of Lorraine and Catalonia conducted by Deleeck et al. (1992), at the Antwerp Centre for Social Policy, assessed not only income and poverty differences, but also the adequacy of social security systems to deal with the economic needs of older persons. For some of these countries, panel data for two waves (l-3 years apart, in the mid- and late 1980's) were examined. Tsakloglou (1994) provided a comparison of both income- and expenditure-based rates, using the 1988 household budget survey data of Eurostat.
The Eurostat studies have several limitations for our purposes. None of the EU comparisons described earlier includes data specifically on older women. The closest any of them approaches this is by looking at older one-person households, which contain predominantly single women (Hagenaars et al. 1994; Deleeck et al. 1992). Another drawback is the absence
32 Social Security Bulletin
of Sweden (until recently), Switzerland, and the United States from the various EU data collection efforts. Finally, the Eurostat data have not been generally available to researchers. In the future, these access limitations are expected to change (Hagenaars et al. 1994, p. 1).
Eurostat has continued work on methodological aspects of the study of poverty, in part through contracts with research institutes in various countries of the Union (Ramprakash 1994). One of its priorities is to capture the multi-dimensional character of poverty (as income, expenditure patterns, self-assessment, and so forth). But because it relies on household budget surveys (the latest dating back to the late 1980's), Eurostat has tended to stress mean expenditure-based measures of poverty. Recent efforts have turned more consistently to income variables and have included a focus on noncash benefits as a component of economic well-being. Most important, a new European Panel on Household Income and Living Conditions, launched in 1994, has a focus on the over-time dimension of relative well-being (Ramprakash 1994).
The European emphasis on a range of different conceptual and methodological approaches in research on poverty has yielded worthwhile findings on several important factors: the use of modified Organization of Economic Cooperation and Development (OECD) equivalence scales (table 2, note l), the overstatement of poverty when surveys do not adjust results by household size, and experience with the statistical matching of household data from different sources (Ramprakash 1994). For the purposes of the present comparative review, Deleeck et al. (1992) and Hagenaars et al. (1994) are included as representative studies.
Luxembourg Income Study
Independently, from 1983, the Luxembourg Income Study began to build an extensive data base offering opportunities for comparative assessments of economic well-being in a larger number of countries. A wave of income data from the early 1990's, covering some of the countries included in the present review, is now becoming available. A long line of research papers based on these standardized microdata have shown their applicability to income-related research questions. There have also been larger scale syntheses, probing key methodological questions of cross-national comparison in depth, as well as more general concerns. Mitchell (199 l), in particular, was able to draw conclusions on issues such as measurement of inequality, comparing income transfer systems, and poverty reduction by means of a thorough analysis of the early LIS files.
From the many LIS-inspired research efforts, we refer here to three other contributions in particular (Buhmann et al. 1988; Smeeding et al. 1993; Whiteford and Kennedy 1995) which included an examination of household incomes among single elderly women in at least five of our comparison countries. Because these studies took into account differences in living arrangements and gender, they set benchmarks for the present review (tables 1 and 2).
Vol. 59, No. 3 Fall 1996
Whiteford and Kennedy (1995) greatly refined the comparative use of LIS data on a broad front. Their research constitutes the most thorough effort of comparative elderly income analysis to date. They devised a more comprehensive "final income" measure that included the effects of indirect taxes and benefits in kind (health and education); used a range...