WILLIAM L. PARISH, Editor MARY C- BRUSTTON, PETER HĽDSTROM, and FRED KNISS, Associate Editors WENDY GRISWOLD and ELISABETH S. CLEMENS, Book Review Editors PAULA S. BRUSH, BRUCE G. CARRUTHERS, MARTHA VAN HAITSMA, CORINNE LALLY-BENEDETTO, and STUART MICHAELS, Associate Book Review Editors MARK JACOBS, Editorial Manager JANE MATHER, Editorial Assistant Consulting Editors ELIJAH ANDERSON ■ RICHARD T. CAMPBELL • MARY L. FEN-NELL ■ NEIL FLIGSTEIN • DAVID F. GREENBERG • CHARLES N. HALABY • CAROL A. HEIMER • LARRY W. ISAAC ■ ROBERT ALUN JONES • LESTER R. KURTZ • ROSS L. MATSUEDA • ANN S. ORLOFF ■ TROND PETERSEN • WALTER W. POWELL ■ JONATHAN RIEDER ■ RACHEL A. ROSENFELD • MICHAEL E. SOBEL • AAGE B. S0RENSEN • MARK C. STAFFORD ■ RONNIE STEINBERG ■ JOHN D. STEPHENS • IVAN SZELENYI ■ EDWARD A. T1RYAKIAN • MARK TRAUGOTT ■ GAYE TUCH-MAN ■ LINDA J. WAITE • CAROL A. B. WARREN ■ SUSAN COTTS WATKINS • DAVID ZARET International Consulting Editore HANS-PETER BLOSSFELD (West Germany) ■ RAYMOND BOUDON (France) ■ PIERRE BOURDIEU (France) ■ MONICA BOYD (Canada) ■ MARTIN BULMER (England) - FRANCIS G. CASTLES (Australia) ■ BONNIE H. ERICSSON (Canada) ■ BERT KLANDERMANS (The Netherlands) ■ BRYAN S. TURNER (The Netherlands) Editorial Bonrd C. ARNOLD ANDERSON ■ GARY S. BECKER ■ RICHARD G. BIERNACKI • CHARLES E. BIDWELL - DONALD J. BOGUE ■ TERRY NICHOLS CLARK ■ JAMES S. COLEMAN -JOHN L. COMAROFF ■ LEO A. GOODMAN ■ PHILIP M. HÄUSER ■ MORRIS J ANO WITZ ■ EVELYN KITAGAWA ■ NANCY L AND ALE ■ EDWARD O. LAUMANN ■ DONALD N. LEVINE ■ DOUGLAS S. MASSEY ■ JOHN F. PADGETT ■ EDWARD SHILS ■ GEORGE STEINMETZ ■ FRED L. STRODTBECK ■ GERALD D. SUTTLES - MARTA TIENDA " WILLIAM J. WILSON The American Journal of Sociology (ISSN 0002-9602) is published bimonthly in July, September, November, January, March, and May by The University of Chicago at The University of Chicago Press, Journals Division, P.O. Box 37005, Chicago, Illinois 60637. Subscription rates, U.S.A.: institutions, 1 year .$64.00; individuals, 1 year $32.00. Student subscription rate, U.S.A.: 1 year $23.00 (photocopy of valid student ID must accompany subscription). ASA and BSA members, 1 year $27.00. All other countries add $8.00 for each year's subscription to cover postage. Subscription agent for Japan; Kinokuniya Company, Ltd-Single copy rates: institutions $11.00, individuals $5.50. Back issues available from 1982 (vol. 88). Volumes in microfilm available from University Microfilms International, 300 North Zeeb Road, Ann Arbor, Michigan 48106; and Kraus Reprint and Periodicals, Route 100, Miliwood, New York 10S46. Subscriptions are payable in advance. Subscriptions will be entered to start with the first issue to be published after receipt of order. Claims for misting numbers should be made within the month following the regular month of publication. The publisher expects to supply missing numbers free only when losses have been sustained in transit and when the reserve stock will permit. Postmaster: Send address changes to The University of Chicago Press, Journals Division, P.O. Box 37005, Chicago, Illinois 60637. Copying beyond Fair Use. The code on the first page of an article in this journal indicates the copyright owner's consent that copies of the article may be made beyond those permitted by Section 107 or 108 of the U.S. Copyright Law provided that copies are made only for personal or internal use or for the personal or internal use of specific clients and provided that the copier pay the stated per-copy fee through the Copyright Clearance Center, Inc. Operations Center, 27 Congress St., Salem, Massachusetts 01970. To request permission for other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collecüve works, or for resale, kindly write to Permissions Department, The University of Chicago Press, 5801 S. Ellis Ave., Chicago, Illinois 60637. If no code appears on the first page of an article, permission to reprint may be obtained only from the author. Second-class postage paid at Chicago, Illinois, and at additional mailing offices. © 1588 by The University of Chicago. All rights reserved. American Journal of Sociology Volume 94 No. 3 November 1988 CONTENTS 465 A General Theory of Expropriative Crime: An Evoli Ecological Approach Lawrence E. Cohen and Richard Machálek 502 Social Networks and Collective Action: A Theory of the Mass. Ill Gerald Marwell, Pamela E. Oliver, and Ralph Prahl 535 Collective Sanctions and the Creation of Prisoner's Dilemma Norms Douglas D. Heckathorn 563 A Theory of Marriage Timing Valerie Kincade Oppenheimer 592 Suburbanization and Segregation in U.S. Metropolitan Areas Douglas S. Massey and Nancy A. Denton j> Commentary and Debate 627 White-Collar Crime, Competition, and Capitalism: Comment on Coleman John Braithwaite 632 Competition and the Structure of Industrial Society: Reply to Braithwaite James William Coleman Review Essays 637 Cultural Analysis and Social Theory Jonathan H. Turner 644 Sociological Theory as Public Philosophy R. Stephen Warner American Journal of Sociology Oliver, Pamela, and Gerald Marwell. 1988. "The Paradox of Group Size in Collective Action: A Theory of the Critical Mass. II." American Sociological Review 53:1-8. Olson, Mancur. 1965. The Logic of Collective Action: Public Goods and the Theory of Groups. Cambridge: Harvard University Press. Opp, Karl-Deiter. 1982. "The Evolutionary Emergence of Norms." British Journal of Social Psychology 21:139-49. Rapoport, A., and A. M. Chammah. 1965. Prisoner's Dilemma. Ann Arbor: University of Michigan Press. Rapoport, Anatol. 1983. Mathematical Models in the Social and Behavioral Sciences New York: Wiley, Rapoport, David C. 1984. "Fear and Trembling: Terrorism in Three Religious Traditions." American Political Science Review 78:658-77. Rothschild-Whitt, Joyce. 1979. "The Collectivist Organization: An Alternative, to Rational Bureaucratic Models." American Sociological Review 44:509-27. Schellenberg, James A. 1982. The Science of Conflict. New York: Oxford University Press. Schotter, Andrew. 1981. The Economic Theory of Social Institutions. New York: Cambridge University Press. Scott, John Finley. 1971. Internalization of Norms: A Sociological Theory of Moral Commitment. Englewood Cliffs, N.J.; Prentice-Hall. Scott, Robert A. 1976. "Deviance, Sanctions, and Social Integration in Small-Scale Societies." Social Forces 54:604-20. Slavin, Robert E. 1983. "When Does Cooperative Learning Increase Student Achievement?" Psychological Bulletin 96:429-45. Taylor, Michael. 1982. Community, Anarchy, and Liberty. Cambridge: Cambridge University Press. Ulimann-Margalit, Edna. 1977. The Emergence of Norms. Oxford: Clarendon. Wehr, Paul. 1979. Conflict Regulation. Boulder: Westview Press. Yakir, Pyotr. 1972. A Childhood in Prison. New York: Macmillan. Yamagishi, Toshio. 1.986. "The Provision of a Sanctioning System as a Public Good." Journal of Personality and Social Psychology 51:110-16. 562 A Theory of Marriage Timing Valerie Kincade Oppenheimer University of California, Los Angeles With the use of a modified job-search theory, a conceptual framework is developed to show that some factors influence marriage timing by either facilitating or impeding assortative mating. Transition-to-work is emphasized because work structures people's lifestyles and is the major source of socioeconomic status; however, its nature is often unpredictable in early adulthood, while other personal attributes emerge early. The theory is applied to the dynamics of assortative mating under two contrasting scenarios: when gender roles are highly segregated, and when women's economic roles start to resemble those of men. Finally, the implications of the analysis for Becker's reduced-gains-to-marriage argument are assessed. In this article, I argue that trends and differentials in marriage timing result, in part, from variations in the degrees of difficulty people encounter in mating assortatively. I explicate a conceptual framework that links assortative mating to the transition to adult economic roles and, in the process, develop a theory of how sex differentials in marriage timing will change as married women's market work becomes more extensive. As a concomitant, I challenge the popular notion that a decline in gains to marriage owing to the increase in women's economic independence is the preeminent factor in the recent rise in delayed marriage. Although this is not often explicitly recognized, a satisfactory marital match can actually be achieved in two different but complementary ways, In the first, a good match is formed through some sort of selection process. Individuals are chosen so that they match on similar or com- 1 This paper is a revision of a presentation given at the 1987 meetings of the Population Association of America in Chicago. I would like to express my appreciation to the many colleagues who took the time to make extensive and valuable critical comments on earlier versions of the paper. In addition to the anonymous reviewers for AJS, I am particularly grateful for the reactions of Richard Berk, Andrew Cherlin, Patricia Gwartney-Gibbs, Margaret Marini, Karen Mason, Mark Plant, Kenneth Sokoloff, and Linda Waite. Requests for reprints should be sent to Valerie K. Oppenheimer, Department of Sociology, University of California, Los Angeles, California 90024. © 1988 by The University of Chicago. All rights reserved. OO02-96O2/89/9403-0OO4$O1.5O AJS Volume 94 Number 3 (November 1988): 563-91 563 American Journal of Sociology plementary traits that they {or sometimes their parents) both value.2 In the second, adaptive socialization during courtship or after marriage modifies the existing traits of one or both partners in order to improve the quality of the match achieved via the selection process alone. Such socialization is one possible solution to the uncertainties that plague all marriages, especially those of relatively young people. In some societies, it may even play an especially important role in household stability.3 While both these processes, selection and socialization, will usually be :: operative, the relative importance of each can vary. For example, if j postmarital socialization becomes more difficult to achieve, then a greater ! reliance must be placed on the marital selection process. I will focus primarily on the factors affecting assortative mating via the selection process. However, I will also pay some attention to the effect of recent socioeconomic changes on the relative importance of these two mechanisms for achieving a good match and the implications of this for marriage timing and marital instability. One approach to investigating the marriage-timing implications of variations in the difficulties of mating assortatively is represented by the "marriage-squeeze" research of demographers. This focuses on the marriage-market implications of variations in the age-sex composition of a population, primarily in response to fluctuations in fertility (Akers 1967; Scboen 1981, 1983). While this approach has important advantages, it is rather limited in the range of assortative problems investigated. For example, it typically seeks to explain only those marriage-timing changes '. that have opposite effects on the two sexes and, therefore, has not been ;. very useful in explaining increases or decreases in age at marriage that are i exhibited by both men and women simultaneously. In this paper, I use a different strategy for investigating the effect of difficulties in mating assortatively on marriage timing. This is to adapt several ideas from job-search theory. The applicability of this theory to marriage searches stems from its explication of the relationship between the closeness of the match sought and the length of time a person must search. When applied to marriage markets, the theory argues for a strong relationship between marriage timing and the problems associated with assortative mating. I will develop this theme and also expand on the 2 There is an extensive, though heterogeneous, sociological literature on assortative mating, much of which is descriptive. See, e.g., Hollingshead 1950; Kennedy 1952; Goode 1964, pp. 332-37; Winch 1955; Riser 1968; Burgess and Wallin 1943; Rockwell 1976; Elder 1969; Glenn, Ross, and Rully 1974; Taylor and Glenn 1976; Udry 1977; Chase 1975; Rubin 1968; Pavalko and Nager 1968. 3 For example, one function of early marriage for traditional Indian and Chinese women was to complete a girl's socialization in the joint households of her husband rather than in her parents' household (Levy 1949). 564 Marriage important differences between searching in marriage markets and job markets. The idea of applying job-search theory to mate selection is by no means new, though work in this area has not yet been extensively developed. Becker and his colleagues have incorporated certain aspects of job-search theory into their discussions of marriage formation (Becker, Landes, and Michael 1977; Becker 1981, chap. 10). However, they did not really use search theory to study the marriage-formation process itself but primarily as a tool to help explain marital dissolution. Moreover, Becker's other work on assortative mating is not developed within a search-theoretic framework but assumes perfect and costless knowledge on the part of marriage-market participants (Berk and Berk 1983). Nor has it been particularly focused on the study of marriage timing (Becker 1974, 1981, chap. 4). In the present article, however, I specifically address the issue of the relevance of search models to marriage timing and, in addition, examine their implications as the division of labor by sex changes. I begin with a brief description of job-search theory, followed by a discussion of the similarities and differences in searching in job and marriage markets. The search model will then be used, in capsule form, to develop the following arguments: 1. Assortative mating is hindered (and hence marriages delayed or disrupted early in their careers) by a relatively high degree of uncertainty about the important attributes that people attempt to match. Moreover, since marriages are planned to last, the long-run benefits of a particular match are substantially affected by the future attributes as well as by the current ones of the persons involved, and there is even greater uncertainty about the nature of an individual's future characteristics. As a consequence, exogenous factors that affect the degree of uncertainty about both future and current attributes will influence marriage timing. 2. The marriage-delaying effects of greater uncertainty can be partially offset by relying on postmarital adaptive socialization to compensate for earlier imperfections in predictions. However, if postmarital socialization becomes a less feasible mechanism for improving the quality of a match, then the role of maritai selection expands, as does the effect of factors affecting the level of early uncertainty about important attributes and life circumstances. In any event, a greater reliance on the selection process to achieve a good match should lead to a later age at marriage. 3. A major source of uncertainty in an industrial society lies in the nature of adult economic roles and in the timing of the transition to a stable work career because work has such a profound influence in structuring a couple's life-style and determining its socioeconomic status. Hence, factors that affect the timing of the transition to a stable work role should also affect marriage timing. 565 American Journal of Sociology 4, If the timing of the transition to adult work roles does have an effect on assortative mating, then highly differentiated gender roles will foster sex differences in age at marriage. Moreover, the nature of the assortative mating process changes when the adult economic roles of men and women start to converge. In particular, there is an increase in the level of uncertainty regarding young women's long-run attributes and a decrease in the role of postmarital socialization as a mechanism for achieving a good match. As a result, women's age at marriage can be expected to rise. SCOPE OF MODEL The theoretical model explicated here is limited in scope. It does not seek to develop a "complete" explanation of marriage-formation behavior or to develop an exhaustive list of all (or most of) the major variables affecting marriage timing. There are just too many random elements involved in searching for a mate for us ever to achieve a full explanation of the marriage behavior of any group or individual. For this reason, the rolfjof socioeconomic factors in assortative mating is emphasized—not because these are necessarily the most important from the individual's point of view but because they are particularly sensitive to exogenous influences and because a person's future socioeconomic characteristics arc often unpredictable at a young age. On the other hand, since they are empirically less systematically problematic, factors that may have the highest subjective priority for those involved may still have little importance in producing macro-level changes or differentials in marriage-formation behavior. Physical appearance, basic personality characteristics, sex appeal, religion, ethnicity, and socioeconomic background are all usually observable when or even before young people start to date.4 Hence, socioeconomic factors, as a source of unpredictability, should be particularly decisive in producing trends and differentials in marriage timing resulting from changes in the nature of the search process. ASSUMPTIONS ABOUT THE ROLE OF RATIONAL CHOICE IN BEHAVIOR Sociologists often dismiss economic models because they object to the economist's emphasis on rational choice. However, many supposedly "nonrational" elements have limited value in accounting for social differences and changes in marriage formation because they do not vary sys- 4 However, sex-ratio imbalances for particular ethnic or religious groups can lead to marriage delays and an increase in the frequency of nonmarriage as well as of heterogamy (Blau, Blum, and Schwartz 1982). 566 Marriage tematically over time and space. Thus, while sexual attraction may be very important in the determination of a marriage partner, unless we can establish that it is more important for some social groups than others or in some time periods than others (and why), it will not be a highly useful predictor of trends and differentials in marriage timing, even assuming we could measure it.5 In addition, regardless of what economists think about the role of rational calculation in behavior, this assumption is not essential to the applicability of job-search theory to marriage searches. Wc need not assume that young people, in general, are calculating in their interpersonal relationships or even that enough of them are to produce significant shifts in aggregate-level behavior.6 For example, one can argue that work is often such an important influence in structuring life that considerable career uncertainties will have an effect on attitudes and the ensuing marriage-formation behavior. From the perspective of the individuals involved, this may take the form of feeling "not ready" to marry, of still desiring to search out an "identity," of not being able to take a "reading" on what another person is really like, and so on. It is also hard to fantasize about the nature of a future life with another person when few clues exist about most of the bare essentials of that future. And I suspect that, given women's traditional economic dependence on their husbands, their fantasies have been particularly important in helping them decide whether a relationship has a potentially desirable future. These feelings—highly subjective and often inchoate but, nevertheless, based on certain realities—will, in turn, often make young people resist forming close relationships. However, the fact that people's feelings and behavior may be rather inchoate does not mean that they are not responding to an objective reality (a reality we can measure more easily than the emotions it engenders) or that it is impossible to subject the process to systematic analysis. JOB-SEARCH THEORY: AN OVERVIEW Job-search theory addresses the problem that, because of the heterogeneity of labor demand and supply, both workers and employers lack 5 One could also argue, however, that sexual attraction is a perfectly rational, though perhaps incomplete, basis on which to form a marriage. If sexual relations are important to a person, he or she would be quite irrational to ignore it as a factor in the mate-selection process. 6 However, young people, and especially their parents, may be relatively calculating about the kinds of social situations in which they (or their children) become involved. Once in a "suitable" setting, however, individuals can more safely allow themselves to be spontaneous and unpremeditated in their relationships (Goode 1963, p. 8; Scott 1965). 567 American Journal of Sociology the knowledge necessary to achieve a perfect and instantaneous matching of workers to jobs.7 Fundamental to the theory is the idea that there is a distribution of potential job offers for any given searcher, only a small proportion of which represents a "perfect" match. Hence, a costly search process has to be conducted. Search is costly for various reasons: there are the direct money costs to job search (e.g., transportation) as well as the psychic costs. Moreover, there are also indirect costs—for example, the earnings lost when a person turns down a job that is less than the perfect match in favor of continuing to search (Stephenson 1976; Salop 1973). Therefore, individuals typically do not continue searching until the perfect match is achieved. Instead, the best strategy is to decide on a minimally acceptable match, expressed in terms of a wage—the "reservation" wage; the job seeker then rejects all job offers with wages below this minimum but accepts the first one with a wage at or above it. The higher the reservation wage, however, the smaller the proportion of jobs in the offer distribution that will fall above it and the longer the time spent searching, since the probability of finding an acceptable match in any given unit period of search is low. This forces the job seeker to continue searching over several periods, thereby increasing the length of the unemployment period. How high the reservation wage is set depends on how much the costs of searching are offset by the returns to search in the form of a better match. Ceteris paribus, the greater the returns from searching, the higher the reservation wage and the longer the time spent searching. The reverse is the case when the returns are low. Much the same reasoning is applied to costs—only here, higher costs lead to a lower reservation wage in order to reduce the time spent searching. In sum, the quality of the match, once made, as well as the length of time spent searching, is a function of the reservation wage. Theoretically, this is set at the point where the returns to further searches are just offset by the costs. Because the reservation wage is determined by equating marginal returns and costs, it is sensitive to factors affecting these two components of the process. For example, the amount of time an individual is planning to work at a job will affect his reservation wage because it affects the returns to search—the shorter the work time desired, the less it will pay to engage in a lengthy search process (Addison and Siebert 1979, p. 174; Ehrenberg and Smith 1982, p. 449). With regard to costs, the existence of search subsidies can reduce the cost of search in each period of time. For example, unemployment benefits will partially subsidize the job-search pro- 7 This discussion of job-search theory rests heavily on the exposition in Ehrenberg and Smith (19S2, pp. 445-50); McCall (1970); Lippman and McCall (1976a). 568 Marriage cess so that the individual can hold out. longer for a higher reservation wage, Once unemployment benefits are exhausted, thereby raising search costs, there is a tendency to settle for poorer-paying jobs (Gronau 1971). While higher search costs lead to a lower reservation wage and shorter search periods, sometimes the costs of search are so high that they exceed the possible benefits (or, alternatively, the benefits are so low that they will never exceed the costs of search). When that is true, dropping out of the labor force is a likely response, and the above is one explanation given for the "discouraged worker" phenomenon (McCall 1970; Lippman and McCall 19766). The length of time spent searching will also be affected by the efficiency of the search process. By search efficiency, I mean the likelihood of finding a good match for each unit cost of investment—that is, the direct and indirect (e.g., time) costs invested per period of search. Basically, this is reflected in the degree of dispersion in the "offer" distribution; a greater dispersion means more can be gained by searching but that in each search period there is a lower probability of coming up with a satisfactory offer (Stigler 1961; Gronau 197.1). If there is little dispersion, however, there will be a higher density of potential matches, and there is less advantage to more extensive searches. Two types of determinants of search efficiency are particularly useful to distinguish. One involves the actual number of potential matches that exist, while the other concerns the searcher's knowledge of and access to these potential matches. For example, if employment is expanding, search efficiency improves as the probability that a given employer is recruiting labor is increased (Stigler 1962). The reverse is the case in a recession. In addition, however, in any given density of potential matches, more knowledge will increase the efficiency of the search process since less time is wasted on nonproductive searches. SEARCHING IN MARRIAGE MARKETS It is not such a great intellectual leap from the analysis of the matching of workers to jobs in the labor market to the matching of men and women in marriage markets (Becker et al. 1977; Becker 1974, 1981). Because of imperfect knowledge, both are processes that are carried out under considerable uncertainty, and, in both, searching can be very costly. Similarly, in the search for a mate, as for a job, one could argue that people will tend to set a minimum acceptance level—what would be considered an "acceptable" match, though not the most preferred or "perfect" match. Those who fall below or outside this area of acceptability will usually not be considered. In both types of searches, the length of time spent search- 569 American Journal of Sociology ing is inextricably bound up with the minimally acceptable match the individual sets and closely tied in with the costs and expected benefits from searching. Despite the parallels between the two types of search processes, there are also several important differences—in the processes themselves, as well as in our ability tu observe them. For one thing, there is the problem of detecting whether a search is actually occurring. By definition, the unemployed are those who are looking for work;8 however, in the case of marriage markets, young people start to date in their early to mid-teens, long before it is reasonable to assume they are searching for marital partners. To complicate matters further, searching for a mate (or just a good date, for that matter) is usually combined with other activities-school, work, recreational activities, and so on.9 A person need not even be looking for a spouse but still find one. Given this fundamental ambiguity in marriage-search behavior,10 the best strategy may be not to try to ascertain whether a search is actually occurring but, instead, to focus on measuring what conditions foster or impede successful searches. A second important difference between the analysis of searching in marriage versus that in job markets is that job-search theory is typically phrased in terms of maximizing income, although it is recognized that nonmonetary rewards may also be involved. However, with marriage searches, the situation is more complex. Although one could discuss matching in terms of the individual's maximizing personal socioeconomic status, this is too narrow an approach to take on an a priori basis. There are numerous returns to marriage that defy easy quantification—at least in dollar terms. For example, marriage provides the opportunity for long-run intimacy and emotional support, a companionship that, by involving historical continuity, promises memories of a shared past. It also provides the major setting for having children, the opportunity for regular (and safer) sex, and so on. In sum, the application of a job-search model to marriage searches does not imply that all that are being matched are socioeconomic characteristics. As a consequence, it is probably not highly meaningful to try to operationalize a marriage-market analogue to the reservation wage. s Even here, however, some investigators have argued that, for young people, the distinction between being unemployed and being out of the labor force is not very meaningful (Shishkin 1976; Leon 1981). 9 However, many economists now argue that a considerable amount of job search occurs while people are still employed (Johnson 1978; MacDonald 1982). Hence, unemployment status is increasingly being viewed as a poor indicator of search status. 10 This ambiguity has interesting substantive implications, however. Combining "search" behavior with other activities reduces the time costs of searching and therefore permits a higher minimum acceptance level and more extensive searches. 570 Marriage By and large, job-search theory is not yet really developmental in character, and herein lies one of its major drawbacks as a tool for analyzing marriage-market behavior. Economists do permit a searcher to learn about the offer distribution in one search period so that, in subsequent searches, he may modify his reservation wage (Lippman and McCall 1976Ö, pp. 173 ff.; Salop 1973). They also sometimes allow for interperiod shifts in the offer distribution, such as those associated with vagaries in the business cycle (Gronau 1971; Kiefer and Neumann 1979; Stephenson 1976; Salop 1973; Gera and Hasan 1982). However, economic models typically do not permit the offer distribution to shift systematically as the individual ages, and they tend to treat the individual's general human capital skills (what he or she is matched on) as known by the individual and fixed from one search period to another.n Nor do they permit the anticipation of future circumstances to affect current search decisions (Lippman and McCall 1976a, p. 158). However, with marriage markets, these assumptions are particularly unrealistic, and the incorporation of age into the model is dictated by several factors. First, in marriage markets, the shape of the offer distribution changes dramatically with age and, with it, the efficiency of the search process. Thus, assuming certain age preferences, the availability of potential mates varies systematically with age as marriage progressively thins out the ranks of the eligible (Goldman, Westoff, and Hammerslough 1984). Moreover, different organizations and institutions will concentrate or disperse the density of potential mates and do so in ways that are related to the individual's age. For example, attending college can greatly increase the efficiency of searching for a mate (Scott 1965). However, this high level of search efficiency may drop off sharply as young adults disperse to more heterogeneous social.milieus. In contrast, armed-forces participation disrupts the marriage markets of males, but this is also highly age related. Second, the degree of uncertainty about important attributes of potential partners or even the searcher's own attributes also shifts systematically with age. This is partly because the uncertainty is not due just to imperfect knowledge of the important existing traits of potential mates; some of the traits that provide essential ingredients in the matching process have not yet been clearly formed and may develop only as adult roles are assumed. Hence, a major difficulty in making long-term matches lies in estimating the nature of important future characteristics on the basis of the incomplete information currently available. And the younger the po- 11 However, there are some exceptions to this in the job-shopping literature (Johnson 1978; MacDonald 1982). 571 American Journal of Sociology tential mates are, the more difficult it is to predict these characteristics— or those of the searcher, for that matter.13 The fact that an early marriage may preclude a premarital matching on adult attributes that have not yet emerged is undoubtedly an important reason for the observed high frequency of "mismatches" and marital instability among earlier marriers (Thornton and Rodgers 1987; Morgan and Rindfuss 1985; Booth and Edwards 1985; Bumpass and Sweet 1972). This suggests that the success of youthful marriages depends partly on how well the partners can predict what their future characteristics and future lives together will be like. The accuracy of these predictions is likely to be influenced by exogenous factors—for example, by the state of the economy. The success of such marriages will also be affected by whether postmarital socialization can compensate for imperfect predictions made during the mate-selection process. Finally, a particularly important difference between job and marriage markets is that, in mate selection, the decision to accept a particular match (whether sought or not) at an earlier time often precludes searching and forming a different and possibly better match later in time. At the very least, it will greatly increase the cost of future searches and also probably reduce the returns. In this sense, there is an opportunity cost to deciding on any particular match, and it should be higher in youth if incomplete information decreases the ability to predict some of a person's important future characteristics.]3>14 Conversely, not accepting an offer (because, e.g., the minimally acceptable match is set too high) risks the opposite opportunity cost. Later matches may not be as desirable as the one refused earlier. This risk is probably greater for women, given their tendency to marry men older than themselves. As a result, the supply of potential mates decreases with age for women but increases for men (Goldman et al. 1984). In sum, finding a mate is, in part, a function of the relative numbers and dispersion of available members of the opposite sex. However, people do not wish to marry just anyone—they want to mate assortativeíy. 12 This might be termed a "maturity" factor, but maturity has a distinct psychological connotation, and lam not talking just about emotional maturity. See also Morgan and Rindfuss's (1985) discussion of "emotional" vs. "social" maturity. 13 If nothing else, a marriage and divorce take time to happen—meanwhile the nature of the marriage market is shifting rapidly. Thus one will never step back into the same marriage market one left to marry—and for some their marriage-market positions may be decidedly worse the next time around; for others they will be better. 14 For this reason also, the job-shopping approach developed by economists has limited applicability to the marriage-search process since "trading up" is usually less of an option. However, the increasing ease with which divorces can be obtained has been reducing the opportunity cost of making a poor match. Besides, we do not really know how much trading up has occurred in recent years. 572 Marriage Hence, if and when a mate is found is also related to the available supply of "suitable" partners, and the success of the search depends not only on numbers but also on the reliability of information about important characteristics of both the searcher and potential partners. The availability of members of the opposite sex, their suitability, and knowledge about their suitability vary systematically with age but not necessarily in the same direction. While information about some important attributes is relatively ascertainable at any time—for example, about age, religion, ethnicity, physical appearance, basic personality features, and so forth—other information is less certain when the person is young but increases with age as adult roles are assumed. However, the availability of potential partners tends to decrease with age as people marry off. Hence, the optimal time for marriage, in the sense of the existence of the greatest number of unmarried persons with some known essential characteristics, is at a relatively young age. The optimum, however, in the sense of the greater availability of information on several important assortative mating attributes, is often at an older age. This age-related aspect of knowledge availability should be especially characteristic of societies undergoing rapid economic change, which, as a result, can lead to greater uncertainties regarding young people's future socioeconomic characteristics. Thus, there is a constant tension between these often-opposing forces in the attempt to optimize the matching process, a tension that may often but not necessarily always be resolved by a considerable amount of modification of attributes via later adaptive socialization. THE TRANSITION TO WORK AND MARRIAGE TIMING Following a long demographic tradition, I argue here that a promising strategy for analyzing marriage timing would be to focus on the timing of the transition to adult economic roles, the nature of these roles in late adolescence and early adulthood, and the various factors that affect them. An individual's current labor-market position affects his or her current ability to marry because it affects the ability to set up an independent household. Hence, economic independence enables already-formed matches to proceed to the marital stage. In addition, however, the assortative mating process itself will be affected by the transition-to-work process and its timing. First, attempts to make matches when young men's current economic position is poor may be discouraged because of the undesirability of long engagements. Second, work provides the socioeconomic means for achieving any given long-run socioeconomic status. However, young people's uncertainty about what kind of work they will be engaged in during their mature adult years makes it difficult to estimate what long-run socioeconomic position accompanies any given po- 573 American Journal of Sociology tential match. More important, perhaps, work structures life in many ways, and this is true whether the individual defines it as the central focus of his being or as simply a necessary evil. Hence, if the nature of adult work roles appears very uncertain, so does the fabric of one's future life. For example, will work involve traveling extensively, reducing time spent together at home? Will the family have to move often because of job transfers? Does the job entail frequent overtime work on nights and weekends or working night shifts? Does it require heavy entertainment responsibilities or involve other types of "two-person career" activities (Papanek 1973)? Is it a high-pressure job that affects the quality of family life? Is it dangerous? Does it involve working with and socializing with interesting or tiresome people? And so on. Thus, by structuring life, work roles impose considerable adult socialization not only on the individual worker but also on those close to him, who are necessarily affected by the spouse's or parent's work role. A person who is a long way from making the transition to a fairly stable work career is therefore very much an -unknown quantity as a potential mate. The result is that factors that affect the timing of this transition should also lead to changes and differentials in the age at marriage. If the nature of the transition to adult economic roles is significant in marriage timing, this immediately raises the issue of gender differences in economic roles. How have the differences in men's and women's traditional economic roles affected sex differences in the timing of marriage? And what have been the effects of women's increasing economic activity outside the family? To answer these questions, I will develop the argument so that differences in gender roles, as well as changes in these differences, are specifically taken into account. THE OPERATION OF MARRIAGE MARKETS WHEN GENDER ROLES ARE HIGHLY DIFFERENTIATED I begin with the oversimplified family situation in which market work is limited to men, while women specialize in home production. Later on, this assumption will be relaxed to see what difference it makes if women do work and are, moreover, expected to work for varying lengths of time throughout the marriage.15 There are a number of reasons for starting with the assumption of highly differentiated sex roles. From an empirical point of view, extensive market work by married women has only developed in the postwar era, so that starting with the assumption of no market 13 I wiil also postpone until then a discussion of the "wealth" women may bring to the marriage, because in modern times this has often been composed, in part, of savings from female earnings. 574 Marriage work and then gradually relaxing it roughly parallel the actual course of historical change (Oppenheimer 1970). From a theoretical point of view, an important sociological literature has elaborated on the functional importance of sex-segregated roles and emphasized the significance of women's "traditional" roles of mothers and homemakers. This is perhaps best typified by the work of Parsons, who argued that sex-role segregation is a functional necessity for marital stability in our society and even for the viability of the society itself.1G This is so, Parsons maintained, because sex-role segregation is the most important mechanism preventing disruptive competition between husband and wife (Parsons 1949).17 This leitmotiv of the significance of gender-differentiated economic roles for marital stability is being echoed strongly in the recent work of the "new home economists" (Becker 1981). Becker basically views unmarried men and women as potential trading partners. A couple marries (trades) because each partner has more to gain by marrying (trading) than by remaining single {not trading). As in all trading relationships, the gains to marriage are based on each person's having something different to trade. In the case of a woman, her comparative advantage in home production leads her to specialize in that, while the man specializes in market work. According to Becker, it is this specialization and the mutual dependence it produces between the sexes that provide the major gains to marriage for each partner. As a consequence, Becker argues, positive assortative mating occurs for traits that are complements—for example, education, intelligence, attractiveness, and so forth—while "negative assortative mating would be optimal for [traits that are] substitutes such as wage earning power" (Becker et al. 1977, p. 1146; see also Becker 1974, 1981). Hence, men with high earnings potential marry women who have low earnings potential but are otherwise superior.18 Becker goes on to argue that "the gain from marriage is reduced by a rise in the earnings and labor force participation of women and by a fall in fertility because a sexual division of labor becomes less advantageous" (Becker 1981, p. 248). Presumably, this is a major factor in the recent rise in the age at marriage and is definitely thought by Becker and his colleagues to be important in the considerable increase in the divorce rate (Becker 1981, p. 248; Becker et al. 1977). 16 For example, in his 1949 description and analysis of the sex-segregated roles of the period, Parsons remarked: "It is scarcely conceivable that the main lines of the present situation could be altered without consequences fatal to the total of our unique society" (Parsons 1949, p. 268). ,r For a more extensive discussion of these issues, see Oppenheimer (1977, 1982) and D'Amico (1983). 1B Becker does not tell us, however, how likely it is to find such women—superior in every way but with few marketable skills and little interest in acquiring them! 575 American Journal of Sociology In sum, Parsons stresses the necessity of minimizing sources of marital conflict in order to maintain marital stability, while Becker emphasizes the importance of the gains to marriage in their effects on marriage formation as well as on marital stability. However, they both see highly differentiated gender roles as central to marriage as an institution. As a consequence, for both, the growth of extensive employment of women is bound to have a major negative influence on the family in our society. Moreover, the view that women's growing economic independence is the major factor in the rise in delayed marriage and marital instability is a common refrain in the demographic literature, expressed by economists and sociologists alike (Ross and Sawhill 1975; Cherlin 1979, 1981, pp. Si ff.; Preston and Richards 1975; Waite and Spitze 1981; Fuchs 1983; Espenshadc 1985; Goldscheider and Waite 1986; Farley 1980). MARRIAGE-MARKET DYNAMICS AND MEN'S AGE AT MARRIAGE Although there arc many factors involved in assortative mating, if the husband is to be the sole or major provider of market goods for the family, the marital socioeconomic status and life-style of a woman will be largely a function of the long-term socioeconomic characteristics and behavior of the man she marries. But since a young man's future prospects are often highly uncertain, this introduces a corresponding uncertainty into the mate-selection process, an uncertainty that will be highly related to career-cycle stage. This uncertainty, and its variations in degree in response to exogenous factors, should affect a young man's short-term marriage-market position because it increases the difficulty of assortative mating. A young man's own desire to marry will also be affected by his career-cycle stage because an early marriage may threaten the completion of important training or prevent job experimentation (Furstenberg 1974; Rapaport 1964). In addition, if he expects upward mobility, he may wish to wait until his socioeconomic characteristics can signal this more clearly and thereby expand his marriage-market options. Moreover, if there is considerable uncertainty about his future life-style, a man may also be uncertain about the kind of woman with whom he would be the happiest. If the costs of search are very high for young men and the returns uncertain, search theory suggests two possible courses of action. One is that high search costs lead to a reduction in the minimally acceptable match, thereby promoting an earlier age at marriage but also a higher probability of a mismatch. Becker has adopted this argument and uses it to help explain marital instability (Becker et al. 1977). However, as I have argued, this strategy exacts a much higher opportunity cost in mar- 576 Marriage riage markets than in job markets since accepting a poorer marital match now may preclude forming a much better match later. If the high costs of searching (or the low returns) are only temporary, then most young people may prefer to postpone seriously searching for a spouse until enough information is available to permit a more productive search process. In short, the analogy to the discouraged-worker argument seems even more appropriate, though, here, it is not necessarily a "dropout" phenomenon per se but really a nonsearch status that is at issue. Moreover, if premarital sex is socially acceptable, this reduces the opportunity costs of not searching for a spouse before a young man's career has stabilized and, with it, his marriage-market position. Hence, the recent transformation of our sexual mores has undoubtedly had a major role in delaying marriage because it has reduced the sexual penalties of such delays—for both men and women. As a consequence, it should reduce the effect of sexual attraction on marriage timing, thereby eliminating an important factor precipitating marriages. ,.. / MARRIAGE-MARKET DYNAMICS AND WOMEN'S AGE AT MARRIAGE What sex differences in the age at marriage will search theory predict when women, as traditional wives, engage primarily in home production? Becker argues essentially that women are so specialized that to be productive they have to be married and that this explains their early age at marriage (Becker 1981, p. 77). One problem with this view is that although women have typically married at an earlier age than men, an early age at marriage has not been universally observed. For example, extensive demographic research on preindustrial northwestern Europe indicates that women's age at marriage was quite high, usually in the late 20s (Hajnal 1965). And, in fact, industrialization has brought with it declines in the age at marriage for many European populations (Modell, Furstenberg, and Strong 1978; Hajnal 1953; Wrigley and Schofield 1981, p. 255; Watkins 1981). A more likely explanation is that the effect of sex-role differentiation \ makes women more marriageable at a younger age than men because ■ there is less early uncertainty about the attributes of women that are important in making a match. Her socioeconomic status and ethnicity, as well as the religion a woman brings to a marriage, are largely inherited, and her physical attractiveness and personality characteristics are also observable at a relatively young age. Moreover, many of her home-production skills are learned primarily in the parental home relatively early in life, and her fecundity is at its peak in young adulthood. In short, American Journal of Sociology the level of uncertainty regarding a woman's future attributes is lower at a younger age relative to the uncertainties associated with a young man's future prospects. Hence, they have generally married at a younger age, on average, than men, but whether it is an early marriage has been heavily dependent on young men's socioeconomic position. Moreover, there have traditionally been strong incentives for women to marry relatively early since the availability of eligible males tends to decline sharply as women get older while the competition from younger women increases markedly (Goldman et al. 1984). Consequently, the much greater opportunity costs of delayed marriage for women should make their transition to marriage much faster than for men and lead to more sharply decreasing marriage | prospects with age (Modell et al. 1978). 1 If the male's career-launching status is an important factor "in marriage j timing, then the age difference between spouses should decline, or even ; reverse, for older female marriers. For a young woman of 18 to marry a boy of 14 or 16 or even of her own age is almost unthinkable. When she is 22, an 18- or 20-year-old male is still not usually considered sufficiently "mature." But, by the time she gets to be 25 or 26, men of the same age or even slightly younger are often already established in their work or are well on their way. And when she is in her late 20s or early 30s, many younger males are indistinguishable in their physical, social, and economic characteristics from men her own age or somewhat older. Hence, age differences between spouses should decline as women's age at first marriage increases. Do we observe these patterns? The differential should not be so apparent for all marriages because as women get older they will be increasingly tapping into the marriage market of divorced and, eventually, widowed males. However, the change in the age differential should show up for first marriages, and we do observe them, as figure 1 shows (NCHS 1985). WOMEN'S MARKET WORK AND MARRIAGE MARKETS IN THE TRADITIONAL SETTING Even under the assumption of the strict division of labor between the sexes after marriage, women have historically brought wealth to their marriages in the form of a dowry (Goode 1963; Arensberg and Kimball 1968; Stone 1960-61). Whatever other functions the dowry served, it could offset the disadvantages of a woman's increasing age in the marriage market, thereby facilitating the marriages of women in their mid-203 and older. Furthermore, with the growth of industry, young women have increasingly worked outside the home (often migrating to cities) and have used some of their earnings to provide a dowry—in fact, if not 578 Marriage 4 3 + 2 1 Years Groom n Older u \ All Marriages I I I í I ) I I 1 I i\l 1 I I I 1 í I I I I i I I j First Marriages -2-- -3 -4-i- 18 20 22 24 26 28 30 32 34 36 38 40 42 44 Age of Bride Fig. 1.—Average number of years groom was older than bride, by age of bride: United States, 1983. Source: National Center for Health Statistics. necessarily in name (Alter 1986; Tilly and Scott 1978).iy In the United States, the substantial rise in the labor-market involvement of young single women not in school started early in the 19th century and reached high levels by the mid-20th century (Oppenheimer 1970; Goldin and Sokoloff 1982; Goldin 1983). I have emphasized that marriage formation is highly dependent not only on the young man's current socioeconomic characteristics but aiso on his estimated long-run socioeconomic prospects since marriages are planned to last. With more prolonged schooling and steep job-progression ladders, however, good estimates of a man's future socioeconomic characteristics may be possible before he is capable of supporting a family, particularly in the style to which the young couple has been accustomed. And in a competitive marriage market, any subsidy of marriage will give young people an incentive to marry on the basis of future prospects rather than present realities. Hence, the possibility of young wives' being em- 19 Nor was this such a departure from rural custom since, in certain parts of Europe at least, a tradition of both sons and daughters' leaving home early in order to work as servants on other farms—what has been called "life-cycle service"—predated the start of industrialization by several hundred years (Berkner 1972; Lasiett 1977; Hajnal 1982; Mcintosh 1984). 579 American Journal of Sociology ployed early in marriage, though not regularly thereafter, may actually reduce the age at marriage because it makes unions less dependent on the often temporarily low earnings of the young men.20 Thus, it is only a short step from working as a single woman to working as a young wife before the first child has arrived—provided a reasonably reliable birth-control technology is available. In this way, market work, combined with the ability to control fertility, can make a union possible before a young man is economically independent but after some reliable assessment of his long-run prospects is possible. This is most likely to be the case in periods of economic expansion and was undoubtedly an important factor in the rapid postwar decline in the age at marriage. In a more faltering economy, such as we have been experiencing recently, the career-cycle*pros-pects of young men are more unpredictable. EXOGENOUS FACTORS AFFECTING MARRIAGE TIMING Occupational Type Because young men's current and future socioeconomic characteristics are sensitive to a number of exogenous factors, marriage-timing behavior will be inherently variable—both cross-sectionally and longitudinally. One cross-sectional factor is the type of career a man pursues (Op-penheimer 1982, pp. 147-62; Furstenberg 1974; Rapaport 1964). Higher-level careers often involve more extensive training and early career uncertainties that make it difficult or career threatening for young men to support families. Men going into stable blue-collar careers, however, may establish themselves at relatively earlier ages. Hence, one should expect occupational differences in marriage timing among men, and these, in fact, have been observed {Oppenheimer 1982, chap. 4). One implication of this is that changes in the occupational structure over time—toward a more professionalized labor force, for example—are a factor in the rising age at marriage. These should also be reflected in geographic differences in marriage timing because of sizc-of-place and regional differences in industrial and occupational structure. Young Men's Income Position If the husband's economic role in the family remains of fundamental significance in American society, then the timing of young men's transition to a stabie work career is still a major factor in the age at marriage of women as well as men. This is often ignored because we cannot easily 20 For a discussion of women's employment early in marriage as a functional substitute for a dowry, see Davis (1972). 580 Marriage include the economic position of potential mates in micro-level analyses of women's age at marriage. For this reason, it is essential to study male marriage timing, a topic that has often been neglected owing to the traditional demographic preoccupation with women as the reproducers. When male and female marriage timings are compared, moreover, we see that i , the median age at first marriage has risen substantially for both sexes. For I example, it increased from 20.8 in 1970 to 23.6 in 1987 for women; however, the rise for men was almost as great—from 23.2 in 1970 lo 25.8 in 1987 {U.S. Bureau of the Census 1987, table 3). One could take men's rising age at marriage as further evidence of the declining gains to marriage because of women's growing labor-force participation and the increasing economic independence this permits. However, the deterioration in young men's labor-market position has been so substantial that this seems the more likely explanation of the changes in male marriage timing and perhaps of female as well. As figure 2 shows, young men's relative economic status has been declining since the mid-1960s, at least. Even their absolute income level (in constant dollars) has '; recently gone down. What is still unclear is why these changes have been occurring and what they tell us about young men's career-entry process. Easterlin argues that the recent sharp decline in the relative economic position of young males is because of the entry of baby-boom cohorts into the labor market. Hence, this situation should soon reverse itself as the baby-bust cohorts reach young adulthood (Easterlin 1980, 1978). On the -----; other hand, Oppenheimer argues that, while relative cohort size is important, other factors are operating as well. Part of the decline is due to greater delays in the transition to a stabie work career caused by major shifts toward a more professionalized occupational structure characterized by relatively steep age-earnings profiles; hence, this change is unlikely to reverse itself in the way relative cohort size will (Oppenheimer |~""~' 1982). In addition, decreases in the relative importance of well-paying lifelong manufacturing jobs in the durable goods industries have also probably had a negative influence on the ease and speed and predictability of many young men's career-entry process (Urquhart 1984). Moreover, the decline in the relative income position of young males may also reflect delays in their career-launching process. For example, there may have been an increase in stopgap types of job attachments rather than in low earnings resulting from being in low-paying career-entry positions (Op-penheimer 1983). To evaluate this situation requires a detailed analysis of changes in young men's occupational attachments and employment regularity. Whatever the reasons for these trends in young men's labor-market position, they are occurring. Historically, trends and differentials in women's marriage timing have primarily been a function of young men's 581 American Journal of Sociology Median Income (1988$) 14000 -r 12000 10000- 8000 ■ ■ 6000 4000 2000 -0 - Median Income (1B8QJ) Income Ratio (%) M I I !■■ I 1 I H I M I I I I I I 1 l-t-l-l-l' M-f-H I t-l l-i-l- j70 -60 -50 --40 -30 ■20 ■ 10 -0 50 53 58 59 62 66 68 71 74 77 80 83 86 Year Income Ratio {%) Fig. 2.—Median income of males aged 20-24 and its ratio to the median for males aged 45-54: United States, 1950-86. Source: U.S. Bureau of the Census, CPR, P-60. ; / - economic characteristics and other factors affecting their availability rather than women's economic characteristics affecting men's marriage timing. This is indicated by the strong correlation between women's marriage rates and factors directly affecting males—business cycles and wars, for example (Rodgers and Thornton 1985). Therefore, it is rather premature to conclude that the current rapid rise in age at marriage of men and women alike is mainly attributable to women's growing economic independence. Some or even most of the changes are tied into changes in young men's economic position—just as they have always been. CHANGES IN WOMEN'S ECONOMIC ROLES AND THE OPERATION OF MARRIAGE MARKETS What happens when married women are expected to have a fairly extensive labor-market involvement throughout their lives instead of limiting it to the period before the birth of the first child? A popular explanation in the social demographic and economic literature is that it increases women's economic independence from marriage, thereby leading to greater marriage delays, a rising incidence of nonmarriage, and a growth in marital instability, not to mention the fertility effects. However, the application of search theory to the problem indicates that the changes in lifetime labor-market involvement will lead to changing marriage behavior for other reasons as well—-some of which are unrelated to the economic-independence effect, though others will interact with it. Hence, the independence effect is probably being given more credit than it is due, 582 '■ ''■'-'' '■"" ''' Marriage • ■■■-• ,-■ •>.- .. - . í , , . ! especially given the persisting strong differentials in female/male earnings.21 Search theory indicates, first, that as women's labor-market attachments have grown, schooling has become more important, as indicated by the substantial rise;in women's college enrollments. As a consequence, the opportunity costs of dropping out of school before completion or of missing other training opportunities have been rising for women. Yet a "premature" commitment to a marriage may require just such dropout behavior. Second, the nature of the matching process changes as women's work involvement becomes more lifelong, although the same search and matching mechanisms are operative. Previously, the husband's market work was the major factor determining the couple's life-style. This required adaptations by both partners to the constraints imposed by the husband's career. Now, there are often two work careers, or potential work careers, that could make possibly conflicting demands that require adaptations. As a result, the feasibility of using postmarital socialization as a corrective matching mechanism is declining, placing more of the burden on assortative mate selection for producing a good match. This will lead inevitably to a rise in age at marriage and greater marital instability since early uncertainties about what is being matched are increasingly characteristic of females as well as males. Serious marriage searches may then be postponed until the emergent nature of each person's attributes and desired life-style is more manifest. However, the length of these periods of uncertainty will vary among socioeconomic groups and according to the nature of the career cycle, as well as because of such exogenous circumstances as the state of the economy. The increasing prevalence of cohabitation among younger people may also represent one type of response to both the declining importance of postmarital socialization as a matchmg mechanism and the increasing uncertainties at an earlier age. Cohabitation gets young people out of high-cost search activities during a period of social immaturity but without incurring what are, for many, the penalties of either heterosexual isolation or promiscuity, and it often offers many of the benefits of marriage, including the pooling of resources and the economies of scale that living together provide. It also facilitates the kind of interaction that increases the knowledge of oneself and of a potential marriage partner and of the kind of mutual adaptations that are so essential to stable 21 For example, the earnings of full-time year-round female workers have remained close to 60% of males'earnings since 1956. However, between 19S0 and 1986 it did rise from 60.5% to 65% (England and Farkas 1986, p. 163; U.S. Bureau of the Census 1987, table 1). / 583 American Journal of Sociology relationships. It can therefore serve as an important prelude to marriage itself—as an extension of the courtship process (Cherlin 1981, pp. 14 ff.). However, cohabitation also provides some of the advantages of remaining single. While it may currently tie people up (though not as much as marriage), its iniluence on future mating behavior is much less, and the long-run financial obligations are also relatively low. In short, cohabitation can be viewed as one type of adjustment to delays in the optimum conditions for assortative matings. Third, paid employment itself now has greater marriage-market functions for women, especially during a period when young men's economic position is increasingly uncertain at the school-leaving ages, thereby reducing the marriage-market efficiency of high schools and colleges.22 For example, work can provide a desirable marriage-market setting or, at least, social networks that extend the perimeter of.the, individual's marriage market. Work also provides the funds for creating an attractive image and for the leisure activities that further enlarge the effective boundaries of the marriage market. If work becomes increasingly important in serving these marriage-market functions, then obviously the nature of the work, the "suitability" of the work setting, the kinds of contacts it can offer, and the money it brings in all become extremely important to young women—whether or not they are set on lifelong "careers."23 And if the marriage-market functions of work are increasing in importance while those of schools are decreasing, despite later school- > leaving ages, then age at marriage will inevitably rise. - ^ ^'- 'l'-^"' í ^-^ We may also incorporate some of the economic-independence effects into the search model. For one thing, greater economic independence subsidizes searching in marriage markets and reduces the economic penalties associated with nonmarriage. This probably encourages an increase in risk taking by some young women in the form of setting higher minimum levels of acceptability for a prospective spouse. This, in turn, will lead to later marriages and increase the chances of never marrying since the older a woman is, the smaller the pool of more eligible males. 22 Furthermore, the marriage-market function of universities has shifted over time. In the postwar period, increases in the proportion of young men and women attending college improved the marriage-market potential of colleges and universities. However, for women, the marriage-market efficiency of colleges has declined with the convergence in the proportions of young men and women attending universities. Thus the sex ratio of those aged 14-34 enrolled in college dropped from 161 in 1964 to 103 in 1984 (U.S. Bureau of the Census 1985, p. 6). 23 In fact, more extensive schooling and employment in a relatively high-level career (though not necessarily at a high level within it) may reduce some kinds of search costs because of greater access to and improved knowledge about a more select group of men as well as the prematching that has occurred because alJ will have certain types of training and work experiences in common. Marriage However, the growing prevalance of divorce and of delayed marriage, for both sexes, reduces this risk for new cohorts entering marriage markets since the supply of suitable mates is disappearing at a slower rate, thereby decreasing the penalties of delay. Work may encourage marriage delays not only because it reduces the dependency of women on husbands but also because it increases their independence of parents'as weil, thereby decreasing the latter's ability to exert effective pressures on their daugh- / ters both to marry and to marry at an early age. ' Women's greater economic independence will be another factor reducing the importance of adaptive socialization in achieving a good match. When women's market work is of short duration and subordinate to their husbands', the woman's willingness to adapt her life-style to the circumstances dictated by her husband's career has long been an essential part of the marriage bargain.24 After marriage, her continued willingness to fulfill this bargain has frequently been assured by the power differential the wife's economic dependence produces. However, as women's attachments to market work increase in strength and duration, not only do the adaptive demands made on married couples change, but so does the bargaining position of each partner. By virtue of her greater economic independence there is an increase in the ability of the wife to bargain for a greater priority being given to her life-style aspirations. The result should be a growing emphasis on premarital matching in the achievement of a good match at the expense of postmarital socialization as the latter becomes a less effective and more potentially conflict-ridden mechanism for " developing a good match. This, in turn, discourages early marriages because of the greater difficulty of achieving good matches at a time characterized by so much uncertainty regarding important future attributes. Marriages are also likely to become more brittle if postmarital socialization is a less influential factor in improving the quality of a \ match, thereby increasing marital instability. During this transformation of women's economic roles in our society, greater delays in marriage and higher rates of marital instability may be observed than will be the case once the situation has stabilized. This is partly because, in a rapidly changing world, some women will have developed serious work attachments only after marriage, leading to a desire to renegotiate their original marriage bargain, a renegotiation that may not always be feasible. In addition, since the possibility of a major work attachment is so new for women, early indications of their future characteristics, as well as their early accomplishments, are more likely to be discounted than are men's at similar career-cycle stages. This, in turn, u I am using the term "marriage bargain" in the sense Willard Waller did many years ago (Waller 193S). /^■•-:; ' "■■' ' ■ :' '!: '' ■ ■'" ''!'v' '' '■■■''-■'-■■■■':■-. t 585 A.U !,■■;'' , i ■ «' .!>, .''••.- .- . ' . - .; American Journal of Sociology weakens the marriage bargaining position of such women, if they want to use their work plans as a bargaining chip, until they have achieved a more established or irrevocable work identity, and, even then, there is often a tendency to treat a woman's market work as a temporary expediency or dalliance rather than as a more permanent commitment. Moreover, the discounting of women's plans and early accomplishments increases the risks an early marriage incurs because of the low priority given to her work goals. CONCLUSION This paper has utilized a modified search-theoretic framework to argue that the timing of the transition to a stable work career has an important impact on marriage timing. If so, then gender-differentiated economic roles will, through the assortative mating process, promote sex differences in marriage timing. The reasons for this are the paramount importance of men's economic role in the family, combined with the varying degrees of youthful uncertainty about the nature of their adult work roles. Moreover, there are both cross-sectional and temporal variations in the timing of the transition to these roles, depending on a variety of factors— occupation, business-cycle conditions, secular economic change, wars, and so forth. For the traditional woman, on the other hand, there is much less life cycle-related uncertainty about her future attributes. One result of these sex differentials in the age pattern of uncertainty is the commonly observed later age at marriage for males. However, the age at marriage for both sexes will be heavily dependent on the timing of young men's entry into relatively stable occupational careers. And as long as men's economic role in the family remains of considerable importance, the nature of the economic prospects of young males should continue to be a major factor in the marriage timing of females, as well as of males, despite changes in women's labor-market behavior. Changes in gender roles will lead to corresponding changes in the age at marriage in part because a woman's more extensive labor-market attachment (whether it is desired or not) adds another set of exogenous factors requiring adaptations by both husband and wife. This, in turn, increases the importance of marital selection over postmarital socialization in fostering a good match. It also raises the level of early uncertainty about women's attributes so that it more closely resembles that of males. Both \ these changes should lead to marriage delays. In sum, all these factors can result in variations in the age at marriage and, in particular, foster the recent trends toward delayed marriage and greater marital instability. But this is due to their effect on assortative mating and is quite aside from the issue of women's growing economic 586 Marriage independence and its possible effect on the gains to the marriage. However, I have also discussed here the influence of the independence effect on the assortative mating process and, through this, raised additional questions about the reduced-gains-to-marriage argument as the major explanation for recent trends in marriage behavior. The issue turns, in part, on the important distinction between whether women's growing economic independence is reducing, or even eliminating, the gains to marriage in general or whether it is primarily the gains to some marriages that have declined? The first alternative suggests that marriage, as an institution, is on the decline, and perhaps even about to disappear, particularly if sex differentials in earnings decrease markedly. After all, we could reason that if, despite the persistence of substantial earnings differentials, the gains to marriage have already declined so much that marked changes in marriage behavior have occurred, what will happen if the earnings differentia! is greatly reduced or even disappears? Taken to the extreme, this argument has quite unattractive implications as it tends to push people into one of two polar positions. If marriage, as a social institution, is deemed important for the stability of society and the reproduction and socialization of the next generation, then the price of women's growing economic independence will increasingly be seen as far too high since it will signify a major decline in our society. Here we are pushed into the position of the New Right, On the other hand, if marriage is viewed as the major instrument for the oppression of women, then its demise via the expanding economic role of women will either be celebrated or, at least, considered a regrettable but unavoidable price for women's liberation. Is it really necessary to be forced into one or the other of these polar alternatives? In this paper, I argue that it is not—partly because there are nonindependence reasons for the changes, but also because a search-theoretic framework reveals how greater economic independence can reduce the gains to some marriages without necessarily having an effect on the gains to marriage in general. This is because greater independence allows women to set a higher standard for the minimally acceptable match—that is, they need not be forced to settle for a poor-quality match or to remain in it despite considerable unhappiness. The consequence is an increase in delayed marriage with some accompanying greater risk of nonmarriage, as well as higher marital instability. But all this is consistent with continued high gains to marriage as well as with a continued desire to marry.25 What it suggests, however, is an increase in the "fric- 23 For example, in the 1980 Study of American Families, Thornton and Frccdman found that only 33% of the 18-year-old daughters would be bothered "a great deal" by not marrying and another 34% would be bothered "some." However, 97% did expect to get married (Thornton and Freedman 1982). 587 American Journal of Sociology tion" with which the family system functions rather than its disintegration. In this way, the approach developed here provides a number of less apocalyptic and more theoretically and empirically challenging alternatives to the reduced-gains-to-marriage hypothesis as an explanation of recent trends in marriage behavior. REFERENCES Addison, John T., and W. Stanley Siebert. 1979. The Market for Labor: An Analytical Treatment. Santa Monica, Calif,; Goodyear. Akers, Donald S. 1967. "On Measuring the Marriage Squeeze." Demography 4:907- 24. Alter, George. 1986. "Households and the Life Course: Natives and Migrants in a Nineteenth Century Belgian City." Paper read at the 1986 annual meetings of the Population Association of America., San Francisco. Arcnsbcrg, Conrad M., and Solon T. Kimball. 1968. Family and Community in Ireland. Cambridge: Harvard University Press. Becker, Gary S. 1974. "A Theory of Marriage." Pp. 299-344 in Economics of the Family: Marriage, Children, and Human Capital, edited by Theodore W. Schultz. Chicago: University of Chicago Press. —-----, 1981. A Treatise on the Family. Cambridge; Harvard University Press. Becker, Gary S., Elisabeth M. Landes, and Robert T. Michael. 1977. "An Economic Analysis of Marital Instability." Journal of Political Economy 85:1141-87. Berk, Richard A., and Sarah Fenstermaker Berk. 1983. "Supply-Side Sociology of the Family: The Challenge of the New Home Economics." Annual Review of Sociology 9:375-95. Berkner, Lutz. 1972. "The Stem Family and the Developmental Cycle of the Peasant Household: An 18th Century Austrian Example." Pp. 34-58 in The American Family in Social-Historical Perspective, edited by Michael Gordon. New York: St. Martin's. Blau, Peter M., Terry C. Blum, and Joseph E. Schwartz. 1982. "Heterogeneity and Intermarriage." American Sociological Review 47;45--62. Booth, Alan, and John N. Edwards. 1985. "Age at Marriage and Marital Instability." Journal of Marriage and the Family 47:67-75. Bumpass, Larry, and James A. Sweet. 1972. "Differentials in Marital Instability: 1970." American Sociological Review ,57:754—66. Burgess, Ernest W., and Paul Wallin. 1943. "Homogamy in Social Characteristics." American Journal of Sociology 49:117-24. Chase, Ivan D. 1975. "A Comparison of Men's and Women's Intergenerational Mobility in the United States." American Sociological Review 40:483-505. Cherlin, Andrew J. 1979. "Work Life and Marital Dissolution." Chap. 9 in Divorce and Separation: Contexts, Causes and Consequences, edited by George Levinger and Oliver C. Moles. New York: Basic. ———. 1981. Marriage, Divorce, Remarriage. Cambridge: Harvard University Press. D'Amico, Ronald. 1983. "Status Maintenance or Status Competition? Wife's Relative Wages as a Determinant of Labor Supply and Market Instability." Social Forces 61:1186-1205. Davis, Kingsley. 1972. "The American Family in Relation to Demographic Change." Pp. 237-65 in Demographic and Social Aspects of Population Growth, vol. 1. Edited by Charles F. Westoff and Robert Parke, Jr. Research Reports of the Commission on Population Groivth and the American Future. Washington, D.C.: Government Printing Office. 588 Marriage Easterlin, Richard A. 1978. "What Will 1984 Be Like? Socioeconomic Implications of < Recent Twists in Age Structure." Demography 15:397-432. ——. 1980. Birth and Fortune. New York: Basic. Ehrenberg, Ronald G., and Robert S. Smith. 1982. Modern Labor Economics: Theory and Public Policy. Glenview, III.: Scott, Forcsman. Elder, Glen H., Jr. 1969. "Appearance and Education in Marriage Mobility."American Sociological Review 34:519—33. England, Paula, and George Farkas. 1986. Household, Employment, and Gender. New York: Aldine. Espenshade, Thomas J. 1985. "Marriage Trends in America; Estimates, Implications, and Underlying Causes." Population and Development Review 11:193-245. Farley, Reynolds. 1988. "After the Starting Line—Blacks and Women in an Uphill Race." Demography (in press). Fuchs, Victor. 1983. How We Live: An Economic Perspective on Americans from Birth to Death. Cambridge: Harvard University Press. Furstenberg, Frank F., Jr. 1974. "Work Experience and Family Life." Pp. 341-60 in Work and the Quality of Life, edited by James OToole. Cambridge: MIT Press. Gera, Surendra, and Abrar Hasan. 1982. "More on the Returns to Job Search: A Test of Two Models." Review of Economics and Statistics 64:151—56. Glenn, Norval D., Adreain A. Ross, and Judy Corder Tully. 1974. "Patterns of Jntergenerationai Mobility of Females through Marriage." American Sociological Review 39:683-99. Goldin, Claudia. 1983. "The Changing Economic Role of Women: A Quantitative Approach." Journal of Interdisciplinary History 4:707-33. Goldin, Claudia, and Kenneth Sokoloff. 1982. "Women, Children and Industrialization in the Early Republic: Evidence from the Manufacturing Censuses." Journal of Economic History 42:741-74. Goldman, Noreen, Charles F. Westoff, and Charles Hammerslough. 1984. "Demography of the Marriage Market in the United States." Population Index 50:5-26. Goldscheider, Frances Kobrin, and Linda J. Waite. 1986. "Sex Differences in the Entry into Marriage." American Journal of Sociology 92:91—109. Goode, William J. 1963. World Revolution and Family Patterns, New York: Free Press. ---------. 1964. The Family. Englewood Cliffs, N.J.: Prentice-Hall. Gronau, Reuben. 1971. "Information and Frictional Unemployment." American Economic Review 61:290-301. Hajnal, John. 1953. "Age at Marriage and Proportions Marrying." Population Studies 7:111-36. ---------. 1965. "European Marriage Patterns in Perspective." Pp. 101-38 in Population in History, edited by D. V. Glass and D. E. C. Eversley. Chicago: Aldine. ——. 1982. "Two Kinds of Premdustrial Household Formation System." Population and Development Review 3:449-94. Hollingshead, August B. 1950. "Cultural Factors in the Selection of Marriage Mates." American Sociological Review 15:6.19-27. Johnson, William R. 1978. "A Theory of Job Shopping." Quarterly Journal of Economics 2:261-78. Kennedy, Ruby Jo Reeves. 1952. "Single or Triple Melting-Pot? Intermarriage in New Haven, 1870-1950." American Journal of Sociology 58:56-59- Kiefer, Nicholas M., and George R. Neumann. 1979. "An Empirical Job-Search Model, with a Test of the Constant Reservation-Wage Hypothesis." Journal of Political Economy 87:89-107. Kiser, Clyde V. 1968. "Assortative Mating by Educational Attainment in Relation to Fertility." Eugenics Quarterly 15:98-112. 589 American Journal of Sociology Laslett, Peter. 1977. "Characteristics of the Western Family Considered over Time." Journal of Family History 2:89-110. Leon, Carol Boyd. 1981. "The Employment-Population Ratio: Its Value in Labor Force Analysis." Monthly Labor Review 104:36-45. Levy, Marion J., Jr. 1949. The Family Revolution in Modern China. Cambridge: Cambridge University Press. Lippman, Steven A., and John J. McCall. 1976a. "The Economics of Job Search: A Survey (Part I)." Economic Inquiry 14:155-89. ——. 19766. "The Economics of Job Search: A Survey (Part II)." Economic Inquiry 14:347-68. McCall, John J. 1970. "Economics of Information and job Search." Quarterly J oumal of Economics 84:113—26. MacDonald, Glenn M. 1982. "A Market Equilibrium Theory of Job Assignment and Sequential Accumulation of Information." American Economic Review 5:1038-55. Mcintosh, Marjorie K. 1984. "Servants and the Household Unit in an Elizabethan English Community." Journal of Family History 9:3-23. ■ ' Modeil, John. 1980. "Normative Aspects of American Marriage Timing since World War II." Journal of Family History 5:210-34. '■■ Modell, John, Frank F. Furstenberg, Jr., and Douglas Strong, 1978. "The Timing of Marriage in the Transition to Adulthood: Continuity and Change, 1860-1975." Pp. S120-S150 in Turning Points: Historical and Sociological Essays on the Family, edited by John Demos and Sarane Spence Boocock. Chicago: University of Chicago Press. Morgan, S. Philip, and Ronald R. Rindfuss, 1985. "Marital Disruption: Structural and Temporal Dimensions. "American Journal of Sociology 90:1055-77. National Center for Health Statistics. 1985. Unpublished marriage data for 1983. Oppenheimer, Valerie Kincadc. 1970. The Female Labor Force in the United States: Demographic and Economic Factors Determining Its Growth and Changing Composition, Population Monograph Series, No. 5. Berkeley: University of California, Institute of International Studies. -----—■. 1977. "The Sociology of Women's Economic Role in the Family." American Sociological Review 42:387-406. ---------. 1982. Work and the Family: A Study in Social Demography. New York: Academic. —-—. 1983. "Men's Labor Market Position and Age at Marriage." Paper presented at the annual meetings of the Population Association of America, Pittsburgh, April. Papanek, Hannah. 1973. "Men, Women, and Work: Reflections on the Two-Person Career." Pp. 90-110 in Changing Women in a Changing Society, edited by Joan Huber. Chicago; University of Chicago Press. Parsons, Talcott. 1949. "The Social Structure of the Family." Pp. 173-201 in The Family: Its Function and Destiny, edited by Ruth Anshen. New York: Harper. Pavalko, Ronald M., and Norma Nager. 1968. "Contingencies of Marriage to High-Status Men." Social Forces 46:523-31. Preston, Samuel H.t and Alan Thomas Richards. 1975. "The Influence of Women's Work Opportunities on Marriage Rates." Demography 12:209—22. Rapaport, Robert N. 1964. "The Male Occupation in Relation to His Decision to Marry." Acta Sociologica 8:68-82. Rockwell, Richard C. 1976. "Historical Trends and Variations in Educational Homogamy." Journal of Marriage and the Family 38:83-95. " Rodgers, Willard, and Arland Thornton. 1985. "Changing Patterns of First Marriage in the United States." Demography 22:265-79. Ross, Heather L., and Isabel V. Sawhill. 1975. Time of Transition: The Growth of Families Headed by Women. Washington, D.C.: Urban Institute. 590 Marriage Rubin, Zick. 1968. "Do American Women Marry Up?" American Sociological Review 33:750-60. Salop, S. C. 1973. "Systematic Job Search and Unemployment." Review of Economic Studies 40:101-201. Schoen, Robert. 1981. "The Harmonic Mean as the Basis of a Reaiistic Two-Sex Marriage Model." Demography 18:201-16. —-----. 1983. "Measuring the Tightness of a Marriage Squeeze." Demography 20:61- 78. Scott, John Finley. 1965. "The American College Sorority; Its Role in Class and Ethnic Endogamy." American Sociological Review 30:514-2 7. Shiskin, Julius. 1976. "Employment and Unemployment: The Doughnut or the Hole?" Monthly Labor Review 99:2-10. Stephenson, Stanley P., Jr. 1976. "The Economics of Youth Job Search Behavior." Review of Economics and Statistics 58:104—111. Stigler, George J. 1961. "The Economics of Information." Journal of Political Economy 69:213-25. —-----. 1962. "Information in the Labor Market." Journal of Political Economy 7:94- 104. Stone, Lawrence. 1960-61. "Marriage among the English Nobility." Comparative Studies in Society and History 3:182-206. Taylor, Patricia Ann, and Norval D. Glenn. 1976. "The Utility of Education and Attractiveness for Females' Status Attainment through Marriage." American Sociological Review 41:484-98. Thornton, Arland, and Deborah Freeman. 1982. "Changing Attitudes toward Marriage and Single Life." Family Planning Perspectives 14:297-303. Thornton, Arland, and William Rodgers. 1987. "The Influence of Individual and Historical Time on Marital Dissolution." Demography 24:1-22, Tilly, Louise A., and Joan Scott. 1978. Women, Work, and Family. New York: Holt, Rinehart & Winston. Udry, J. Richard. 1977. "The Importance of Being Beautiful: A Reexamination and Racial Comparison." American J oumal of Sociology 83:154-60. U.S. Bureau of the Census. 1965—1986, Current Population Reports. Series P-60, 1965-1986. —------. 1967. Trends in the Income, of Families and Persons in the United States: 1946-1964. Technical Paper No. 17. Washington, D.C.: Government Printing Office. —------. 1985. "School Enrollment—-Social and Economic Characteristics of Students: October 1984 (Advance Report)." Current Population Reports. Series P-20, No. 404. ———. 1987. "Households, Families, Marital Status, and Living Arrangements March 1987 (Advance Report)." Current Population Reports. Series P-20, No. 417. Urquhart, Michael. 1984. "The Employment Shift to Services: Where Did It Come From?" Monthly Labor Review 107:15-22. Waite, Linda, and Glenna D. Spitze. 1981. "Young Women's Transition to Marriage." Demography 18:681-94. Waller, Willard. 1938. The Family: A Dynamic Interpretation. New York: Cordon. Watkins, Susan Cotts. 1981. "Regional Patterns of Nuptiality in Europe, 1870-1960." Population Studies 35:199-215, Winch, Robert F. 1955. "The Theory of Complementary Needs in Mate Selection: A Test of One Kind of Complementariness." American Sociological Review 20:52-56. Wrigley, E. A., and R. S. Schofield. 1981. The Population History of England, 1541- 1871. Cambridge: Harvard University Press. 591