Variations on an Equitable Theme: Explaining International Same-Sex Partner Recognition Laws M. V. Lee Badgett Associate Professor of Economics University of Massachusetts Amherst, MA 01003 USA lbadgett@econs.umass.edu and University of Amsterdam (2003-4) For presentation at the conference on “Same-sex couples, same-sex partnerships, and homosexual marriages” Stockholm University September 2003 Please do not quote or cite this paper without the author’s permission. Comments welcome! Introduction As an historic political and policy innovation, the formal legal recognition of same-sex couples has spread rapidly throughout Western Europe since 1989. The Netherlands and Belgium now allow same-sex partners to marry, and Norway, Sweden, Denmark, France, Germany, and several other countries have created a new partner recognition status alongside marriage for same-sex couples and sometimes for different-sex couples (see Table 1). Other countries in Europe have recognized same-sex partnerships for particular purposes, such as for immigration rights, or in geographically limited jurisdictions. Much more limited change has occurred outside Europe. In North America, Canada will soon expand marriage for same-sex couples from Ontario to the entire country. But decentralization in family policy and powerful political opposition have made the United States much slower to take action, with only a few states providing even a limited form of recognition of same-sex couples.1 These cross-national legal innovations in family policy hold much promise for improving the lives of lesbian, gay, and bisexual people who might enter same-sex partnerships, strengthening the economic and social positions of their families. Furthermore, such policies are an example of institutional change that varies across countries, presenting an opportunity for studies that may shed light on the underlying causes of change. In my overall research project, I seek to understand both aspects of this historical shift in the European context. First, why have some countries adopted same-sex partner recognition (SSPR) laws while other countries have not? Second, what impact will such policies have on the status and well-being of same-sex couples’ families? I will use both qualitative and quantitative methods to analyze the determinants of passage of SSPR laws and their impact on families. In this introductory paper, though, I explore more specifically the first question, asking whether variations in social norms about homosexuality, sexuality, and gender or other factors are likely candidates for explaining whether a country has adopted an SSPR. Literature on SSPR laws To date, most work on this topic has been done by legal scholars, with most attention going to the mechanics of the laws and relatively little to the reasons for the passage of such laws. Kees Waaldijk posits a useful “law of small change” to account for the passage of samesex partner recognition laws in Europe (Waaldijk, 2001). Over time, he notes, European countries have gradually, steadily, and mostly sequentially liberalized laws that place gay men and lesbians in a second-class position. The standard path is decriminalizing sodomy, equalizing the age of consent for same-sex sexual relationships, enacting anti-discrimination legislation, and finally addressing partnership and parenting rights. The most recent step for Denmark, Norway, Sweden, Iceland, and the Netherlands was to create a near-marriage relationship (or access to actual marriage in the Netherlands and Belgium) for same-sex couples. The expanded cohabitation rights encoded in partner recognition laws in France and Germany represent a less advanced stage in the many possible small changes on the way to full equality. Table 1 summarizes the countries with such laws. 1 The state of Vermont has created “civil unions” that give same-sex couples the same rights as married couples within the state’s borders. California and Hawaii also offer a less expansive form of recognition to same-sex couples, but no other states appear close to recognizing same-sex couples. 1 Eskridge offers a similarly optimistic path for change (Eskridge, 2002). He argues that cohort replacement and intra-cohort shifts in homophobic attitudes create a more liberal social environment for gay and lesbian people. These environmental changes then encourage gay people to be more open about their existence and to mobilize politically. The openness and mobilization provide more information that falsifies the stereotypes and misinformation that perpetuate homophobia, thus contributing to less homophobia and more progressive legal changes that favor equality for gay people and same-sex couples. More progressive legal changes then perpetuate the cycle of change. Waaldijk, Eskridge, and others also note the importance of thinking about broader social characteristics that might contribute to some nations’ more progressive policies for gay couples. Henning Bech ascribes Denmark’s innovative 1989 registration law for same-sex partners to a Danish tradition of “frisind” or broad-mindedness combined with a sense of social obligation to be an innovator in matters of equality (Bech, 1992, p. 142; see also Wintemute, 2001). More generally, same-sex partner recognition laws may have emerged in countries with a weaker Christian or other fundamentalist political presence (Bech, Eskridge), strong gay and lesbian social and political mobilization (Bech; Eskridge), strong traditions of tolerance for minorities that include liberal attitudes toward gender nonconformity and sexuality (Waaldijk; Eskridge), less religiosity (Waaldijk; Wintemute), less direct democratic decision-making institutions (i.e. fewer opportunities to put the issue before voters; Waaldijk), and a declining importance of marriage as an institution (Eskridge; Wintemute). But the small steps taken by any given country actually constitute a set of outcomes of larger political processes set in a particular economic, social, and cultural context. Seeing change as incremental and perhaps even as inevitable does not answer why some countries have started down the incremental path but others have not, nor why some countries have progressed faster than others. One country’s path to change may be another country’s bloody ideological battlefield. As Klawitter and Hammer suggest in their study of the spread of sexual orientation antidiscrimination laws in the United States, a half-way position might be a final consolation prize rather than a clear step in the direction of continued change (Klawitter and Hammer, 1999). Each step toward passage of laws might require increasing political mobilization and might engender increasing symbolic opposition, thus slowing or stopping the process of change in some countries. While the suggested routes to change identify likely factors that may well be linked to the existence of SSPRs, the rather deterministic nature of the line of reasoning offered by legal scholars is overly optimistic, if history is any guide. For instance, John Boswell’s historical work demonstrates that same-sex couples had at least some forms of social and legal recognition in medieval Europe, leaving a long gap of centuries of quite unequal treatment until 1989. A longer historical perspective suggests that legal changes are neither inevitable nor necessarily predictable continuations of current trends. Conceptualizing Institutional Change An alternative way of approaching the question draws on social science theories of institutional change. In this more theoretical context, I would supplement and reorder the influences suggested earlier within a discussion of the behavioral effects of institutions and of institutional change. We can think of SSPR laws as institutions in the sense that they are a “set of rules that structure social interactions in particular ways” (Knight, 1992, p. 2). Rules about 2 who is recognized as a significant other will influence the legal and economic relationships of members of the family as well as the relationship of family members to non-family members and the state. In earlier work, I outline the many ways that lack of access to marriage may influence the behavior and economic status of same-sex couples in the United States (Badgett, 2001), although the behavioral effects are not my interest in this paper. Neoclassical economists tend to see formal (e.g. laws) and informal (e.g. cultural norms) institutions as simultaneously being constraints shaping human behavior and constructs fulfilling important social functions. Institutions evolve to make societies more productive, perhaps in a Darwinian process of survival of the institutions that make for the fittest economies (e.g. North, 1991). In the context of family law, economists and other scholars have suggested several ways that marriage enhances efficiency: • Promoting specialization of labor: Becker argues that the marriage contract allows for increasing household efficiency through a sexual division of labor that promotes higher productivity through lifelong specialization (Becker, 1991). Without the presumed permanence of the household that the marriage bond implies, specialization by either party would not necessarily be efficient in the long-term.2 • Reducing transaction costs: Pollak (1985) argues that marriage promotes efficiency through reducing transaction costs for couples, removing the need to renegotiate the terms of the legal relationship as couples experience changed circumstances.3 • Providing social insurance: Pollak also notes that wealth and income pooling across individuals and families provides insurance against bad times, such as the failure of a harvest or the loss of a job. • Signaling commitment: Eskridge (1996) argues that the willingness to marry is an important signal of commitment to a relationship. The commitment to a long-term relationship underlies the specialization, transaction costs, and social insurance functions of marriage. • Taking advantage of economies of scale: By encouraging households made up of more than one adult, marriage promotes situations in which economies of scale might be achieved, that is, where doubling the inputs of time and other resources results in more than double the output of family-related goods and servivces, such as meals or child development. From these perspectives, the legal institution of marriage promotes efficiency at a social level and at a couple level.4 Both individual couples and societies have an incentive to seek out and utilize this relatively efficient institutions. Other social scientists propose that institutions are less the outcome of an efficient struggle for survival than an outcome of social and political bargaining in which the more powerful are able to shape institutions that serve their own interests (e.g. Knight, 1992). From this perspective, any larger collective value of the institutions is incidental rather than intentional. Much feminist economic analysis shares this emphasis on power and conflict in the structuring of institutions (e.g. Folbre, 1994; Agarwal, 1997; Badgett, Davidson, and Folbre, 2000). In this view, both the structure of institutions and the rules of access to institutions that 2 At least one economist has used this view of economic institutions to argue that the efficiency of existing marriage laws justifies the continuing prohibition of same-sex marriage in Canada. 3 See a related argument for allowing same-sex couples to marry in Badgett (2001). 4 Some may argue that marriage is neither necessary nor sufficient to achieve the efficiencies outlined here. The possibility of multiple “equilibria,” i.e. a variety of equally efficient social institutions, would certainly affect our judgment about the efficiency impact of the creation or destruction of an institution. But for a given institution, the efficiency impact of changing access to those institutions would always remain. 3 confer public benefits will be shaped by political competition at least as much as by economic competition. In thinking about how these two broad perspectives on institutions shed light on the development of SSPR laws, we might well first question how to characterize those policies. Do the SSPR laws represent the sort of monumental change that most discussions of institutions entail, or are they simply a sort of technical correction to existing marriage laws, changing only one of the conditions for entering a marriage and implying no fundamental change in the structure or legal meaning of marriage? Some lesbian, gay, bisexual, and transgender advocates clearly view the fight for marriage as assimilationist—that is, reflecting an acceptance of mainstream heterosexual values--and at best a relatively minor change in the legal landscape that shapes the quality of human life. However, I would argue that the tenacity of many other LGBT advocates and their allies who push for the right to marry and the equal ferocity of the right-wing (and even more moderate) efforts to prevent same-sex marriage suggest that we are talking about a change of historic proportions, both for couples and for society. Once we see SSPRs as examples of new institutions or radically changed old institutions, applying the efficiency and social conflict perspectives give us a sense of why such institutional changes might emerge. To the extent that different countries experience different economic and political contexts, we can possibly explain the differential emergence of SSPRs and shed light on the debate over institutional change. From the broad perspectives, we can distill more specific hypotheses about causal relationships, although these causal factors are not necessarily mutually exclusive. Existing marriage laws may become inefficient and vulnerable to change for many reasons. First, the emergence of a new family form—or at least a newly visible family form—such as gay or lesbian couples might promote change in existing norms (informal institutions) and laws (formal institutions) to accommodate the “newcomers.” Second, the development of state-provided benefits and services might make marriage less necessary to assure the survival and well-being of adults and children.5 The impact of such de-familializing policies (to use Esping-Andersen’s phrase) is ambiguous, however. On one hand, if marriage has less functional value for creating economically efficient family outcomes, then the demand for change by same-sex couples would likely be reduced. On the other hand, to the extent that the state favors married couples (e.g. for inheritance taxes) and as long as other contractual elements of marriage have meaning for the couple (e.g. rules for the division of property when a marriage ends), entering marriage will enhance the economic well-being of individual couples. On a political level, moreover, there might be less resistance to changing marriage laws in such contexts, an effect that also fits into the conflict perspective discussed below. The explicit incentives for marriage to evolve to include gay couples is a bit murky here. If marriage improves a couple’s economic well-being, then perhaps happier and healthier samesex couples and their children would improve the economic performance of a society. Thus simply publicly proclaiming the existence of same-sex couples and their needs should be sufficient to expand the access to marriage to include same-sex couples. From a different perspective, the kind of mechanism hypothesized by Florida and Gates (2001) could fit into an evolutionary perspective. Florida and Gates argue that places characterized by social diversity and tolerance attract talented workers, explaining their finding of a positive correlation between the proportion of same-sex couples in a metropolitan area and that area’s concentration and 5 Esping-Andersen calls this process of state provision of social insurance, child care, and other traditional responsibilities of the family “de-familialization.” (Esping-Andersen, 1999) 4 growth of high-technology industries in the United States. If partnership recognition laws both create and reflect national values promoting sexual or family diversity, then having those laws might disproportionately attract more highly-educated migrants who value diversity. Alternatively, SSPR laws might emerge because of shifts in the context for social conflict, with this possibility adding two more hypotheses, labeled here as third and fourth. In hypothesis three, groups that favor allowing same-sex couples to marry might gain in political strength or social bargaining power. Fourth, perhaps the defined interests and goals of those who maintain political power change. If an influential but relatively conservative group were to decide that they favored allowing same-sex couples to marry, or at least no longer saw such marriages as conflicting with the interests of that group, for instance, then such a change would be more likely to occur. Certain political precursors, such as state provision of family-related goods and services (as noted above), could reduce the resistance to changing marriage laws since marriage provides less of a marker of social status. While this model focuses on predicting change within a country—whether it adopts an SSPR—the model should also help explain the pattern of SSPRs at a point in time. The same hypotheses would apply to asking why some countries have an SSPR in 2003 while others do not. In the next sections, I propose and carry out some simple tests to see which variables (that correspond to the hypotheses) appear to explain the difference between countries that have and do not have such laws. Data and Explanatory Variables In operationalizing measures that address these four specific different hypotheses, the outlines of a rather large research project emerge.6 For purposes of this paper, however, I will focus attention on a preliminary analysis of data from the World Values Survey and other data sources that provide cross-national data. However, I also outline a few thoughts on additional ways of capturing important explanatory factors, many of which I hope to accomplish in the future. The World Values Survey (Inglehart, et al., 2000) is an ambitious project that collects cross-sectional, individual level data on values and norms about many different topics, including sexuality, gender, and homosexuality, in fifty different countries. So far data from three separate surveys is available, with first waves conducted in 1981-84 (24 societies), the second from 199093 (43 societies), and the third from 1995-97. (A fourth wave from 1999-2000 is not yet publicly available.) Survey language and concepts were translated for each country and were administered by professional survey organizations in Western countries and elsewhere mostly by local survey researchers (Inglehart, 1997). Not all questions were asked in each country’s survey. All but a few countries have representative national samples of adults over 18 years old, selected through stratified multistage random sampling. Data was collected in face-to-face interviews. Some countries over-sampled particular subpopulations, such as younger age groups or racial groups, and the WVS provides individual sampling weights used here to account for over-sampling within countries. Sample sizes ranged generally from 1,000 to almost 3,000 in the twenty-seven countries used in this paper. Because the unit of analysis here is the country, I aggregate individual responses up to the country level. 6 My empirical methods will include detailed case studies comparing and contrasting the experiences of several European countries. 5 With the exception of Denmark, all of the countries with SSPR laws enacted those policies after 1990, so here I will focus mainly on measures from the 1990 WVS. All of the countries with nation-wide SSPR laws are European, but for comparison purposes I have culled data for all of the OECD countries represented in the WVS, plus South Africa, which is of interest because of its rapid movement toward legal equality for gay men and lesbians. The OECD collects data from different national sources on its member states, providing a supplement to the WVS. (1) Efficiency: The emergence of visible same-sex couples. Unfortunately, neither the WVS nor the OECD data are helpful with this hypothesis. Some countries might collect data on their own that allow estimating the number or percentage of same-sex couple families or households, although such measures are unlikely to be available on an historical basis. To some extent, gay and lesbian political mobilization may constitute the most important measure of visibility for same-sex couples, but since couples could exist and even be highly visible without being politically mobilized, I take up political factors below in relation to the conflict hypothesis. Although other markers of visibility are more difficult to quantify, same-sex couples will also become more visible as landmark legal cases related to partner rights or to parenting are brought before courts, as corporate marketers seek to attract a gay and lesbian niche market for goods and services, and as governments respond to the HIV epidemic among men who have sex with other men. In future work, I will investigate the extent of these efficiency-related variables using interviews and other qualitative research methods. These measures are not included in this paper’s quantitative exercise. (2) Efficiency: Declining social and economic value of marriage. The lower the value of marriage in a country, the less the existing legal institution contributes to efficiency and the less political resistance is likely to block dramatic change. To measure the economic value of marriage to couples, I use data on various measures. • Women’s labor force participation rate: This measure proxies the degree of individual economic autonomy, especially for women. • Gender wage gap: The percentage difference in the wages of women and men also provide a perspective on the prospects for economic autonomy for individuals in the marketplace (as well as, perhaps, social attitudes about women). • Beliefs about marriage: The World Values Survey, discussed below, asks respondents whether they agree that “marriage is an outdated institution,” providing a general crossnational measure of beliefs about the value of marriage. • State provision of necessary social insurance benefits: To capture the economic need for marriage, I use the percentage of public social expenditure as a percentage of GDP, which captures the extent of state support for individuals and, to some extent, the degree of de-familialization in a country. • Actual choices to marry or not marry: I use OECD data on divorce rates and calculations from the WVS measures of nonmarital cohabitation rates. • Attitudes toward marriage: The WVS asks each respondent, “Do you agree or disagree with the following statement? ‘Marriage is an out-dated institution.’” The measure is the percentage of a country’s respondents who agree and may reflect whether respondents believe that legal marriage still serves important social, economic, or cultural functions. In each case, the conceptual framework implies a positive correlation between the explanatory factor and the existence of an SSPR. 6 (3) Conflict: Shifts in political power and collective resources. The particular factors that are likely to reshape the political landscape to favor SSPRs are the mobilization of a lesbian and gay political movement, the existence of allies among left-wing political parties, the feminist movement, and labor movements, and the salience of religiously motivated political opposition. The WVS provides only one relevant measure of religiosity, which might predict the degree and strength of religious-based opposition to SSPRs: • Religiosity: The WVS asked each respondent: “Apart from weddings, christenings, and funerals, about how often do you attend religious services these days?” The answer options were: More than once a week, once a week, once a month, Only on special holy days, once a year, less often, never. The summary statistic used here is the percentage of respondents in a country who say they attend at least once a month. • Union membership: The data on the percentage of people who are union members comes from the OECD. The extent that union members are allied with left-oriented political parties, and to the extent that those parties’power is enhanced by greater union membership, I expect that the proportion of workers who are union members to be positively related to the existence of an SSPR. To further assess the role of shifts in group power, in the future I hope to conduct interviews with key actors in these different political locations in the case study countries. I also hope to collect data on other measures of political resources, including membership in and budgets of gay political groups, feminist organizations, and conservative Christian organizations. Data on parliamentary representation of legislators from the left and right will also be gathered to assess collective resources. (4) Conflict: Shifts in political interests and social norms about sexuality, gender, and homosexuality. More tolerant social norms and public opinion about sexuality and gender, in general, and homosexuality, in particular, might make institutional change more likely by reducing opposition to change at an individual level for political decision-makers and for voters. The WVS asks many questions about attitudes that could predict more liberal social normsa nd therefore greater political support from voters and legislators for the concept of recognizing same-sex partnerships: • Homosexuality: The WVS has two ways of measuring attitudes about “homosexuality”: o “On this list are various groups of people. Could you please sort out any that you would not like to have as neighbors?” The measure is the percentage of respondents in a country mentioning “Homosexuals.” (This option was not available on the 1981 survey.) o “Please tell me for each of the following statements whether you think it can ever be justified, never be justified, or something in between, using this card.” (Scale from 1—never-- to 10-always) The measure is the country’s average score for the statement, “Homosexuality”. • Gender norms: The WVS asks a number of questions related to gender norms. The two used here capture norms related to children: o “If someone says a child needs a home with both a mother and a father to grow up happily, would you tend to agree or disagree?” The measure is the percentage of respondents in a country who say they agree and might pick up on anxieties that respondents might have about children raised by same-sex couples, which could influence support for an SSPR law. 7 • o “For each of the folowing statements I read out, can you tell me how much you agree with each. Do you agree strongly, agree, disagree, or disagree strongly? ‘A working mother can establish just as warm and secure a relationship with her children as a mother who does not work.’” Sexual freedom: The WVS also asks, “If someone said that individuals should have the chance to enjoy complete sexual freedom without being restricted, would you tend to agree or disagree?” The measure is the percentage of respondents in a country who say that they “tend to agree.” Since “complete sexual freedom” might reasonably include having a same-sex sexual partner, this measure could pick up tolerance for gay people and their relationships. Findings The Appendix presents the WVS-generated figures for each OECD country that was covered by the WVS. In only a few countries is a question asked in all three waves, so measures of change are not systematically available or calculated here. A quick look through the findings shows that the measures occasionally bounce around in odd ways, such as the doubling of frequent church attendance in South Korea from 1981 to 1990 and then a halving back to the 1981 values in 1995. (Such patterns of big fluctuations are rare, however.) However, in most countries with data from two or more waves, attitudes towards homosexuality have become more liberal over time. I propose two simple ways of assessing the contribution of these measures of national norms to the enactment of SSPR laws – comparing the means for groups of countries for each variable and using a simple regression model to predict whether countries have an SSPR. With such a small number of countries, the regression model will not have much power to distinguish the various factors, so I use various parsimonious specifications to look for important correlations. Table 2 compares the average scores for several groups of countries defined by having an SSPR law.7 Comparing the shaded first and third rows shows that countries with SSPR policies often have different scores in the predicted directions from OECD countries without SSPRs. Perhaps a better comparison is between the second row, which presents averages for the European countries that do not have SSPRs, and the third row for European countries with SSPRs, a comparison that controls to some extent for cultural similarities across European countries. The fourth and fifth rows split the SSPR countries into those with quasi- or actual marriage for same-sex couples (Denmark, Finland, Iceland, Netherlands, Norway, Sweden) and those that allow same-sex cohabitors to register for a much more limited set of benefits (France, Germany, and Belgium, which did not allow marriage when this analysis was undertaken). Efficiency-related variables: Most of the efficiency-related variables show little difference between countries with and without SSPRs. In Table 2, we see little difference in women’s labor force participation, public expenditures on social benefits, the divorce rate, and attitudes about marriage. Marriage remains an important modern institution, according to the vast majority of respondents in every country. Only 14% of people in each country, on average, agreed that marriage is outdated. The one striking difference is the percentage of cohabiting 7 Because the country is the unit of analysis, these averages are not weighted by population or sample size. Please exercise caution in interpreting these averages, since the number of countries covered varies across waves, even within rows. Also, Hungary is not included in the SSPR countries since there is no formal registration process. 8 couples. Respondents in the SSPR countries are more than five times more likely to report “living together as married” in those countries. Conflict-related variables: Many more differences emerge between the countries with and without SSPRs in the conflict-related variables. Both the religiosity measures and the union density figures are quite different for the two groups. People in countries with SSPRs attend church less frequently and are more likely to be union members, supporting the idea that conservative religious opposition is less powerful and left-oriented parties more powerful than in non-SSPR countries. Overall, the sexuality and gender-related norms of respondents in SSPR countries are more liberal, also supporting the conflict hypothesis. In non-SSPR countries, 46% of people would not like to have a homosexual neighbor, while only 22% of people in SSPR countries report a similar prejudice. People in SSPR countries also believe that homosexuality is more likely to be justified. The gender attitude measures are a bit different in SSPR countries, with those in countries with the quasi-marital relationships being much less likely to state that children need a mother and father, and more likely to believe that working mothers can have warm relationships with their children. Unexpectedly, people in those same countries are much less likely to agree with complete sexual freedom. Respondents in France and Germany were more like the respondents in non-SSPR countries with respect to attitudes about gender roles and sexual freedom. Multivariate analysis: The regressions reported in Table 3 allow several factors to vary at once. The dependent variable is one if a country has an SSPR law, and zero if it does not. Each column in Table 3 is a separate regression using just the variables with reported coefficients. Each coefficient shows the impact of a change in the independent variable on the probability of a country’s having an SSPR, holding the other variables in the equation constant. For instance, in the first column, the coefficient of –0.012 means that a country with 40% of its respondents saying that they would not like a homosexual neighbor is 12% less likely to have an SSPR than a country where 30% of respondents do not want a homosexual neighbor. The small sample size limits the power of these tests, but the regressions demonstrate, first, that some explanatory factors are correlated with each other, and second, that some explanatory factors appear to be closely related to having an SSPR.8 The regressions in Table 3 test first for the importance of tolerant attitudes about homosexuality. The coefficients on the two attitudes toward homosexuality variables in the first four specifications (i.e. columns) show that tolerance of homosexuality is positively related to having an SSPR. A country’s respondents’ greater willingness to have a homosexual neighbor and stronger belief that homosexuality can be justified are significantly and positively associated with an SSPR. Notably, however, that effect disappears when the variable measuring the rate of unmarried cohabitation is added. Countries with more cohabitors are more likely to have an SSPR. Since cohabitation is positively correlated with positive attitudes about homosexuality (the correlation coefficient is 0.6 and statistically significant with “homosexuality justified”), the impact of tolerant norms becomes insignificant in column (5). Cohabitation continues to have a statistically significant and positive impact no matter what other variables are included in the regression, as revealed by reading along the row for the cohabitation variable. Most other variables are not statistically significant in the bulk of the regressions, with a few exceptions. The public social expenditures variable is positively related and significant in many specifications, but adding back the tolerance of homosexuality variables reduces its 8 Some variables discussed earlier, such as the women’s labor participation rate and the divorce rate, were never statistically significant, so they are left out of the regressions reported in Table 2. 9 significance. In some regressions, the church attendance measure of religiosity is negatively related to having an SSPR, and union membership is sometimes positively related to an SSPR, both effects as expected.9 Oddly, the gender wage gap has a positive and significant relationship to having an SSPR—countries with higher wage gaps are more likely to have an SSPR after holding cohabitation, social expenditures, and tolerance of homosexuality constant. In a simple regression of the wage gap on SSPR, a specification not reported here, higher wage gaps were associated with a decreased likelihood of an SSPR, however. Tentative conclusions The regressions are helpful in identifying correlates of SSPRs. The efficiency-related hypotheses fared less well than the conflict-oriented hypotheses in the regressions, though. Measures of de-familialization, that is the declining dependence on the family and the increasing dependence on either the state or market for certain kinds of economic support, have mixed effects on the likelihood of a same-sex partner recognition law. Neither a higher women’s labor force participation rate nor a lower wage gap increases the likelihood of an SSPR. In most countries, marriage remains a vital institution according to reported attitudes about marriage, but that attitudinal variable does not vary systematically for countries with and without SSPRs. Two other variables clearly support an efficiency-related hypothesis, though. The regressions suggest that countries with SSPRs tend to devote a greater share of GDP to public social expenditures, and higher rates of unmarried cohabitation are also associated with SSPRs. Both of those variables capture the practical and the perceived economic and social value of marriage. Interpreting these findings is complicated, however. Recall that the hypothesized causal links of these “efficiency” related variables are through reduced political resistance to changing the rules of access to a devalued institution, or at least to its rights and responsibilities. But a devalued institution would be less attractive to those who are excluded, as well, unless important benefits of marriage remain. In future work, I will attempt to create measures to better capture the remaining economic significance of marriage for couples. Further investigation is also important for understanding why higher cohabitation rates, in particular, are associated with SSPRs. In some countries cohabitors have achieved access to certain marital rights through a different legal status, perhaps making it easier to extend that more limited status to same-sex couples than in other countries. Or perhaps cohabitors have more political power in countries where they are numerous, giving them the ability to enact their more liberal beliefs in the context of institutional change. The more explicitly conflict-oriented variables tell a fairly straightforward and consistent story in the regressions. People in countries with SSPRs are less likely to be regular church attendees and are more likely to be union members, suggesting a stronger liberal presence and a smaller conservative religious base for opponents of SSPRs. The explicit normative measures also support the conflict hypothesis. People in SSPR countries have more liberal attitudes toward homosexuality, as predicted. One might also interpret the positive effect of public social expenditures as having a normative element, as well, if those higher expenditures occur because of a national goal of greater equality in well-being and that belief in equality is extended to same-sex couples. However, in other regressions not reported here, including the Gini coefficient, an explicit measure of income equality, did not 9 Church attendance and union density are strongly negatively correlated, with a statistically significant negative correlation coefficient of –0.64. Union membership and church membership are not signficantly correlated. 10 have a statistically significant effect on the probability of an SSPR (nor a consistent sign on the coefficient). Although these tests are fairly blunt at this point, the strength of the correlations is surprising and encouraging. Developing better measures of key variables, as well as more research into the underlying processes at work, should help to further clarify why some countries have SSPRs while others do not. Understanding the factors promoting institutional change will also assist policymakers and advocates in countries without SSPR laws in their strategizing about the likely routes to change. 11 References Agarwal, Bina. ""Bargaining and Gender Relations: Within and Beyond the Household," Feminist Economics, Vol. 3, No. 1, 1997. Badgett, M. V. Lee. Money, Myths, and Change: The Economic Lives of Lesbians and Gay Men, University of Chicago Press, Chicago, 2001. Badgett, M. V. Lee, Pamela Davidson, Nancy Folbre, and Jeannette Lim. “Breadwinner Dad, Homemaker Mom: An Interdisciplinary Analysis of Changing Gender Norms in the United States, 1977-1998." Unpublished, 2000. Bech, Henning. "Report from a Rotten State: 'Marriage' and 'Homosexuality' in 'Denmark'", in Modern Homosexualities: Fragments of Lesbian and Gay Experience, ed. by Ken Plummer, Routledge, 1992, 134-147. Eskridge, William N., Jr. Equality Practice: Civil Unions and the Future of Gay Rights, Routledge, New York, 2002. -----. The Case for Same-Sex Marriage. New York: Free Press, 1996. Esping-Andersen, Gosta. Social Foundations of Postindustrial Economics. Oxford University Press, Oxford, 1999. Florida, Richard, and Gary Gates. “Technology and Tolerance: The Importance of Diversity to High-Technology Growth,” The Brookings Institution Survey Series, Center on Urban and Metropolitan Policy, June 2001. Folbre, Nancy. Who Pays for the Kids? Gender and the Structures of Constraint. London: Routledge, 1994. Inglehart, Ronald, et al. World Values Surveys and European Values Surveys, 1981-84, 199093, 1995-97 [computer file and codebook]. ICPSR Version, Ann Arbor, MI: Institute for Social Research [producer], 2000. Ann Arbor, MI: Inter-University Consortium for Political and Social Research [distributor], 2000. Klawitter, Marieka, and Brian Hammer. “Spatial and Temporal Diffusion of Local Antidiscrimination Policies for Sexual Orientation,” in Ellen D. B. Riggle and Barry L. Tadlock, eds., Gays and Lesbians in the Democratic Process, Columbia University Press, New York, 1999. Knight, Jack. Institutions and Social Conflict. Cambridge: Cambridge University Press, 1992. North, Douglass C. “Institutions.” Journal of Economic Perspectives, Vol. 5, No. 1, Winter 1991, pp. 97-112. 12 Pollak, Robert. “A Transaction Cost Approach to Families and Households.” Journal of Economc Literature. Vol. 23, 1985, pp. 581-608. Vaid, Urvashi. Virtual Equality: The Mainstreaming of Gay and Lesbian Liberation. New York: Anchor Books. 1995. Waaldijk, Kees. “Small Change: How the Road to Same-sex Marriage Got Paved in the Netherlands.” In Robert Wintemute and Mads Andenaes, ed., Legal Recognition of Same-sex Partnerships: A Study of National, European and International Law, Hart Publishing, Oxford, Portland, OR, 2001. Pp. 437-464. Wintemute, Robert. “Conclusion.” In Robert Wintemute and Mads Andenaes, ed., Legal Recognition of Same-sex Partnerships: A Study of National, European and International Law, Hart Publishing, Oxford, Portland, OR, 2001, pp. 759-773. Wintemute, Robert, and Mads Andenaes, ed., Legal Recognition of Same-sex Partnerships: A Study of National, European and International Law, Hart Publishing, Oxford, Portland, OR, 2001. 13 Table 1: Countries with national recognition of same-sex partnerships Country (year enacted) MARRIAGE Netherlands Marital rights and responsibilities Marital rights not included All Presumption about legal status of second parent to a child born to a married woman in same-sex couple. Belgium REGISTERED PARTNERS— QUASIMARITAL Almost all Denmark (1989) Norway (1993) Sweden (1994) Iceland (1996) Finland (2001) REGISTERED COHABITATION Liability for debts; common property; joint taxation; housing; insurance (France) Support obligation; Joint tenancy; inheritance; pension and health Germany (2001) insurance; immigration (Germany) France (1999) UNREGIS. PARTNERS Some inheritance and common property rightsuHu; pension rights; housing rights Right to church wedding (some countries); adoption rights; residence and/or nationality requirements; access to artificial insemination; not portable to other countries Inheritance rights; child-related rights; alimony (France) State-supported financial benefits (Germany) Joint adoption, artificial insemination Hungary (1996) Sources: Eskridge (2001); Wintemute (2001) 14 Table 2: Comparisons of measures across country types CONFLICT-RELATED VARIABLES EFFICIENCY-RELATED VARIABLES Country Non SSPR Other Europe All SSPR Regis. Partners Regis. Cohab. Overall average Gender Pub soc Working Sexual Marriage Attend % trade No homo- Homo- Need Women's wage expend Divorce % Cohab outdated mother freedom relig union sexual sexuality mom LFPR gap (%) % GDP rate couples institution serv rel w/kid OK memb neighbor justified & dad 1990 1990 1995 1995 1990 1990 1990 1990 1990 1990 1990 1990 57.6 23.3 20.5 35.2 4.0% 12.7 43.5 18.6% 46.1 3.2 89.2 65.6 28.1 56.9 22.4 23.1 34.3 3.6% 13.7 46.8 13.2% 39.6 3.5 90.2 64.9 31.6 55.8 18.9 24.1 40.7 15.2% 15.9 18.5 23.0% 21.8 4.7 85.9 72.0 25.7 58.4 21.1 22.5 39.0 17.6% 13.4 13.9 28.5% 17.1 5.1 80.4 79.3 18.6 52.6 16.0 26.5 42.7 11.6% 19.5 25.3 12.0% 29.0 4.1 94.2 61.2 36.3 56.9 21.9 21.9 37.1 0.1 14.0 34.3 20.6% 36.0 3.8 87.7 68.3 15 27.1 Table 3: Regression Coefficients, Dependent Variable is Having an SSPR Law Variable Constant Women’s LFPR Gender wage gap (1) 0.78** (0.17) (2) -0.44* (0.24) (3) -0.91* (0.51) (4) -0.394 (0.49) Marriage outdated Cohabitation rate (5) -0.232 (0.41) (6) 0.49 (0.65) 0.049** (0.012) Public social expend (% GDP) 0.013** (0.005) Union membership Church attendance Attitudes: Sexual freedom Working women’s relationship with child Homosexual neighbor not OK Homosexuality justified Adjusted R2 N 0.008 (0.005) -0.009* (0.005) 0.007 (0.007) -0.003 (0.007) -0.001 (0.01) -0.009 (0.01) 0.198** (0.07) 0.33 23 0.199** (0.07) 0.57 20 0.067 (0.06) 0.63 23 0.183** (0.062) 0.63 20 -0.012** (0.004) 0.26 23 0.20** (0.06) 0.29 26 16 Table 3, continued Variable Constant Women’s LFPR Gender wage gap Marriage outdated Cohabitation rate (7) -0.134 (0.09) 0.055** (0.008) Public social expend (% GDP) Church attendance (8) -0.360** (0.17) 0.046** (0.010) 0.014 (0.009) (9) -1.54** (0.65) (10) -1.95** (0.71) (11) -1.68* (0.80) (12) -2.00** (0.85) 0.026* (0.014) 0.027* (0.014) 0.008 (0.013) 0.041** (0.014) 0.050** (0.019) 0.028* (0.016) 0.030* (0.16) 0.049** (0.014) 0.037** (0.017) 0.053** (0.018) 0.038* (0.018) 0.001 (0.0064 0.052** (0.018) 0.035* (0.018) 0.003 (0.004) (13) -0.51 (0.84) (14) -0.30 (0.33) (15) -0.200 (0.62) -0.001 (0.013) 0.043** (0.018) 0.022 (0.18) -0.002 (0.005) -0.001 (0.013) 0.040** (0.013) 0.018* (0.010) -0.003 (0.003) 0.040** (0.013) 0.018 (0.012) -0.003 (0.004) Attitudes: Sexual freedom -0.000 (0.006) Working women’s relationship with child Homosexual neighbor not OK Homosexuality justified Adjusted R2 N Notes: ** stat signif at 5% level -0.001 (0.007) 0.002 (0.008) 0.65 24 0.67 23 0.67 19 0.65 18 0.65 19 0.068 (0.062) 0.65 19 0.62 21 0.63 22 0.62 21 * stat signif at 10% level 17 Appendix: Data on norms from first three waves of the World Values Survey Country Australia Austria Belgium Canada Czech Rep. Denmark Finland France W Germany E Germany Hungary Iceland Ireland Italy Japan Korea Mexico Netherlands Norway Poland Portugal Spain Sweden Switzerland Turkey UK US S Africa No homo Homosexuality justified Marriage outdated Sexual freedom OK Kids need a mom & a dad Attend relig serv neighbor frequently 1990 19951981 1990 1995-98 1981 1990 1995-98 1981 1990 1995-98 1981 1990 1995-98 1981 1990 1995-98 98 24.7 3.8 4.6 13.3 18.5 27.6 30.6 97.1 70.7 24.9 25.1 43.4 3.3 11.9 38.8 93.5 44.0 23.5 3.2 3.9 20.1 22.4 23.1 28.1 86.2 91.8 36.1 30.4 29.7 3.1 4.1 13.0 12.5 19.3 23.6 67.1 77.9 46.5 39.8 53.2 5.7 10.5 11.6 99.2 10.8 11.7 6.1 4.7 24.4 18.0 11.1 11.1 53.3 72.8 7.3 10.8 25.2 32.4 3.4 4.4 4.3 17.1 12.6 20.6 9.0 11.9 26.5 56.5 85.9 72.7 10.7 10.9 11.4 24.4 3.2 3.9 30.7 29.1 30.2 32.2 86.7 93.6 17.7 16.8 33.8 10.2 3.4 4.5 7.1 14.1 14.7 30.0 21.5 36.1 24.7 92.2 94.3 84.7 37.3 33.9 24.3 34.1 16.5 3.9 5.9 14.1 28.2 48.7 23.2 97.0 89.2 20.1 8.8 75.3 1.4 2.7 16.6 11.4 10.9 18.0 98.6 16.7 23.0 20.1 3.3 5.4 12.6 6.3 23.8 23.5 78.1 78.5 10.9 9.4 33.2 3.4 3.1 15.9 9.9 25.1 17.2 70.8 82.6 85.5 87.4 36.8 2.5 3.9 22.3 14.1 27.5 42.1 91.2 96.7 50.4 50.6 68.5 2.5 2.4 3.7 23.7 7.0 10.3 10.0 13.3 17.6 88.7 95.2 93.8 12.8 14.3 10.8 4.2 2.2 1.6 2.1 13.3 15.4 14.5 17.1 90.5 95.7 27.4 59.7 26.4 60.2 39.9 2.3 2.9 2.9 18.9 16.9 25.8 29.9 31.2 23.5 90.8 87.6 76.0 74.9 62.2 65.1 8.4 6.1 7.5 16.7 19.5 27.5 32.3 75.6 74.1 38.7 29.6 19.5 16.2 3.6 4.1 5.7 13.1 10.1 13.5 8.3 10.3 10.6 78.7 85.6 85.3 14.8 12.6 12.4 66.1 1.8 2.8 7.5 10.5 22.2 20.5 98.3 97.0 84.9 73.1 49.6 2.3 23.2 24.2 93.0 41.2 29.6 20.3 2.8 3.9 5.5 24.3 15.4 16.7 28.9 53.4 56.3 87.2 92.6 88.6 53.6 40.6 36.6 17.7 11.2 4.5 4.5 7.0 15.2 14.1 16.4 17.2 22.4 17.2 70.0 85.2 81.8 13.8 10.2 10.7 18.5 4.2 6.5 13.1 24.1 44.1 31.6 88.5 38.4 24.2 91.7 91.9 1.6 11.3 7.7 23.8 29.4 95.7 97.5 37.7 43.5 31.1 21.7 3.5 3.5 5.2 13.1 17.6 22.1 30.8 67.6 74.5 24.6 23.4 38.3 29.5 2.4 3.1 3.7 8.1 8.1 10.6 21.6 26.9 25.4 62.7 73.1 73.6 59.7 57.0 54.7 50.5 2.3 2.3 2.6 19.9 13.2 21.7 23.4 24.2 27.2 83.4 92.2 90.2 60.8 69.7 18