Economic Stress in Lives: Developmental Perspectives

Journal of Social Issues, Vol. 44, No. 4 , 1988, pp. 25-45

Economic Stress in Lives: Developmental Perspectives Glen H. Elder, Jr. Universiry of North Carolina at Chapel Hill

Avshalom Caspi Harvard Universiry

A major task for research on the social costs of economic stress is to trace how macrosocial changes affect increasingly smaller social units and ultimately those microsocial phenomena that directly influence children in theirfamilies. In this paper, we specify linkages between macroeconomic change and children’s development by tracing deprivational effects through family adaptations in the household economy and in personal relationships. Our findings from research on children and families of the Great Depression are discussed in relation to an interactional model of the process by which families adapt to stressful times. The theme of economic stress in lives connects two distinct lines of research that have converged in a perspective on the life course. One involves the study of social and economic change, from macro trends to economic cycles and local plant closings. The other investigates lives and family patterns over time. A merger of these research foci reflects a growing awareness of the mutual dependency of a changing society and changing lives, a dependency that has special meaning in the drama of stressful times, collective and individual. Our goal in this paper is to outline and illustrate an approach to the study of economic stress that links social change to individual development. This paper is based on a program of research on social change in the family and life course (Social Change Project), currently supported by NIMH Grants MH-41327 and MH-41827. Glen H. Elder, Jr. was supported by a NIMH Senior Scientist Award (MH00567). We are indebted to the Institute of Human Development, University of California, Berkeley, for permission to use archival data from the Oakland Growth Study and the Berkeley Guidance Study. Correspondence regarding this article should be addressed to Glen H. Elder, Jr., Department of Sociology, University of North Carolina, Chapel Hill, NC 27599. 25 0022-4537/E8/l200-0025506.00/1 Q 1988 Ihe Society for the Psychological Study of Social Issues

26

Elder end Caspi

Linking Levels of Analysis: Macroeconomic Change and Individual Development Any study of human development in a rapidly changing environment must attend at some point to the principal ways by which processes at different levels of analysis are linked. Linkages may be thought of as overlapping regions of different conceptual systems (Elder, 1973), such as the individual and the interactional context, the small group and the larger social organization, and the organization in relation to its surrounding environment. As an illustration, consider the relationship between social class and children’s cognitive development. How does social class affect children’s thinking? Environmental characteristics associated with different class backgrounds are pervasive and range from the quality of housing conditions, to the nature of toys, to the number of books available to the child. Thus, social class may be related to children’s perceptual, cognitive, and perhaps even motivational development because the physical environment associated with this background provides a substrate of stimulation required for development, independent of the individual’s own behavior and the behavior of others in the environment (Wachs, 1979). Another version of the process linking class background and children’s development might connect the nature of parents’ work with the achievement motivation of children through a specific type of family environment (see Kohn & Schooler, 1983). To bring out the motivational relevance of family patterns, we must identlfy relatively concrete characteristicsof adult achievement models (such as the parents’ personal standards, aspirations, values, and social prestige), and also specify the key aspects of achievement training-the encouragement of self-direction, emotional support, presentation and reinforcement of standards of excellence, etc. Without losing sight of the structural variable, we must specify aspects of the family environment that provide explicit linkages to the achievement motivation of children. The relative effects of each of these linkages can be tested empirically. Unfortunately, our understanding of the linkage between macrosocial phenomena and individual development has been impeded by a general failure to explicate and test theories of how and why these relations occur. An adequate analysis involves three key principles (House, 1981). First, we must adequately understand the nature of the social phenomena in question. As we shall see below, econodc stress almost always has multiple dimensions or components, and we need to be clear about which of these are most relevant for the behavioral outcomes of interest. Second, we must recognize that the effects of social events are transmitted through stimuli that impinge directly on the individual. Thus, a major task for research on the costs of economic stress is to trace how macrosocial processes affect increasingly smaller social structures and ultimately

Economic Stress in Lives

27

those microsocial phenomena that directly impinge on the individual. Finally, we must understand how and to what extent macrosocial phenomena, and the more proximal microsocial phenomena they produce, affect individual development. This article views both family relationships and individual factors as linkages between stressful economic times and lives, with an emphasis on the experience of parents and their children. We begin with some analytical considerations about economic stress and the causal process from the macroeconomic level to the microsetting of deprivational risk. The second part of the article examines the influence of family hardship on children and adults. We explore the implications of a shift toward more labor-intensive operations in hard-pressed families, and then consider the interplay between individual development and family relationships.

Economic Stress and the Causal Process Several definitional issues must be dealt with as a preliminary step in linking economic stress to behavioral outcomes. These include a distinction between change in economic status and absolute level of financial well-being; the specification of multiple routes to economic loss; a dynamic model of the family economy that shifts attention from static concepts, such as income status, to reciprocal effects between family units and economic conditions; and finally, the concept of interdependent lives. As used in this paper, economic stress refers to the pressures and strains that arise from a substantial income loss, in contrast to the circumstances of chronic hardship or poverty. Adverse changes in state tend to be more stressful than chronic deprivations, and the reason has much to do with the accommodation of individual preferences to the environment. Following W. I. Thomas (1909), we think of a transition to greater disadvantage as a traumatic disturbance of habit. Drastic economic change disrupts customary ways of living and behaving, and thereby establishes a painful disjuncture between family claims and the resources with which to achieve these claims (Elder & Caspi, in press). Substantial income losses are surprisingly common in contemporary society. Data from a nationwide panel (the Michigan Panel Study of Income Dynamics: Duncan & Morgan, 1985) reveal that a third of the population experienced income losses of 50% or more between 1969 and 1980, a period of dramatic economic crises (Duncan, 1987). The general picture that emerges from research on family income suggests that income is highly volatile over the life span and that chronic poverty applies to a relatively small group (Bane & Ellwood, 1986). Apart from the undeniable distress of the chronically poor, it appears that the deterioration of one’s economic status often produces more stress than the actual hardship state itself (Campbell, 1981). In what follows we examine the effects of decremental change-in economic status for individual and family functioning.

28

Elder and Caspi

A second definitional issue concerns economic phenomena and routes to change in economic status. Economic stress is generally viewed in terms of unemployment (Kelvin & Jarrett, 1985; Keyssar, 1985). But job loss is concentrated in the lower strata and represents only one of numerous sources of economic decline and hardship. Instead of laying off workers, firms may spread the work by placing all employees on a part-time schedule. A change of this sort can produce losses that exceed a half-week’s pay. Cutbacks in hourly wages or monthly salary, and demotion to a lower pay scale, are other routes to substantial income loss. Family events, especially changes in composition, are often coupled with substantial declines in family income (Duncan, 1984). Separations and desertions represent additional sources of economic hardship. Indeed, losses above 50% of prior income are commonly experienced by divorcees and their children (Weitzman, 1985). With work and family events so highly related to economic adversity, the isolation of these effects remains one of the most challenging analytical tasks in the study of economic change as a cause of behavioral disorder. A third definitional issue concerns the family economy. The typical account of hard times links the events of a failing economy to adverse change in family welfare. But the latter is also the result of changes within the family or household. Household expansion through childbirth or residential “doubling up” increases resource needs, perhaps at a time when family income is declining through the withdrawal of mothers from the work force, the departure of working children from home, or the labor force departure of men through health impairment or retirement (Oppenheimer, 1982). Consistent with these interdependencies, perspectives on the life course view the family system as both a cause and a consequence of economic events and change (Elder, 1978). External economic change affects the internal dynamics and patterns of family life, and the latter, in turn, affect the economic wellbeing of the family. For example, economic misfortune may prompt adaptations within the household economy, including alterations in the social roles of members, and also compositional change, from the postponement of children to the addition of relatives and the departure of older youth. These adaptive responses, in turn, may serve to modify the family’s economic situation, which again influences family decision making. Adolescent childbearing and divorce illustrate this reciprocal process. Birth of a child to an adolescent girl severely limits her chances for the future unless she marries or briefly lives with her parents and benefits from their child-care assistance (Furstenberg, Brooks-Gum, & Morgan, 1987). In the case of divorce and child custody by the mother, children experience a much lower economic level as they enter a one-parent household. Pressures from this situation may lead to residential doubling up with relatives and/or remarriage. An understanding of the family’s financial well-being is thus best achieved by reference to the concept of “family economy,” which views the family as a flexible unit whose eco-

Economic Stress in Lives

29

nomic status depends both on the contributions of all its members as well as on the workings of the outside economy. Finally, studies on change in economic status require a concept of lives lived interdependently (Elder, Caspi, & Burton, 1988). The dynamics of interdependence represent a means by which macroevents and extrafamilial stressors affect individual development, and as we shall see in greater detail, their understanding is a prerequisite for analyzing the social costs of economic stress. The concept of interdependent lives represents a central theme of family systems theories and their expanding influence in developmental psychology (Minuchin, 1985). The basic concept of systems approaches is that the family represents a social group, and its functioning as a whole is qualitatively different from the sum of its parts. This holds true because the properties of the family as a whole are derived from the properties of the relationships between individuals in the family and not just from the characteristics of the individuals as separate persons. Thus, to explicate the nature of the home environment and family dynamics in stressful times requires a framework that links family members and that considers personal relationships in terms of a network of interactions. The Analytic Model Our discussion of economic stress up to this point relates a generalized economic decline to family income loss and family adaptations. Figure 1 places these factors in a larger framework. The causal process extends from the economic downturn to parent and child behavior with select feedback loops. The intervening variables provide a skeletal account of the mediational process by which economic change influences children. Two features of the framework deserve particular attention. First, it draws attention to the options and resources available to the family prior to the occurrence of economic loss. For example, family alternatives in economically deprived circumstances vary according to career stage and household composition, and the age, sex, and number of family members. Similarly, the psychological costs of economic stress depend in part on the attributes (e.g., personality characteristics, social support, physical health) that are brought to the change situation. Second, the model stresses the critical role played by adaptive responses in mediating linkages between economic realities and their influence on children. These responses are the adaptations families use to regain control over desired outcomes in the face of economic change; each one has significant consequences for family members in both the short and long term. The Samples This approach to economic stress stems from a program of research on socioeconomic stress and mental health that began at the Institute of Human

Elder and Caspi Macro-economic Decline

I Resources

Definition of the j,

'.

1

* Economy Relationships *FI

f i

+

Adaptations Individual Explosive Calm Supportive

f.

~

Parent Outcomes Economic recovery Nurturant behavior Consistent Supportive

Continued hardship Hostile Arbitrary, erratic

Child Behavior Fig. 1. Analytic model relating macroeconomic change to family and individual outcomes.

Development, University of California, Berkeley. Our empirical work draws on two longitudinal studies of child development that were launched around 1930 (Eichom, 1981). The Oakland Growth Study is based on an adolescent cohort born in 1920-1921. The Berkeley Guidance Study is based on a cohort of preschool children born in 1928-1929. The Oakland Growth Study began in 1931 when fifth-grade children from five schools in the northeastern sector of Oakland were selected for a projected analysis of mental, social, and physical development in a normal sample of boys and girls. Selection was based on two criteria: willingness to participate and anticipated residential permanence in the area. This procedure produced a sample of 167 children, who were studied continuously from 1932 to 1939. Data were collected annually from multiple sources-teachers, study participants, peers, classmates, and staff observers. Mothers were interviewed on three occasions, in 1932, 1934, and 1936. Participants in the Berkeley Guidance Study were selected for a study of normal development from a cross section of births in the city of Berkeley during an 18-month period in 1928-1929. Most of the children were Caucasian and Protestant, and two-thirds came from middle-class families. All members of the

Economic Stress in Lives

31

sample were assigned to matched groups: an intensively studied group (N = 113), on whom we report in this article, and a less intensively studied group (N = 101). Across the 1930s, annual data on the parents, children, and the family unit as a whole came from observations by Institute field workers, from teachers, and from self-reports (interviews and questionnaires). (Participants in the Oakland and Berkeley studies have also been followed up in three time periods throughout the 1980s.) As it happens, history performed a natural experiment throughout the Great Depression (Elder, 1981). No event could have seemed more improbable to Bay Area residents in 1930 than a prolonged economic crisis, but between 1929 and 1933 the average income of Oakland Growth Study families declined by nearly 40% and that of Berkeley Guidance Study families dropped 29%. Variation in income loss thus serves as our point of departure for examining the effects of economic change on Oakland and Berkeley study members and their families. In Oakland, two deprivational groups were identified according to income loss (1929-1933) relative to decline in cost of living (about 25% over this time period). Families suffered asset losses with some frequency only when economic loss exceeded 40% of 1929 income. Therefore, deprived families were defined in terms of income losses above 35%. All other families were categorized as nondeprived. This division proved equally appropriate for the Berkeley sample. By this criterion, 36% of the Berkeley middle-class families were economically deprived, compared to 57% of the Berkeley working-class families. Deprived families were more prevalent in both strata of the Oakland cohort (56 of middle class vs. 69% of working class) than in the Berkeley cohort, a difference that partially reflects the more commercial-industrial character of that city’s economy. With pre-Depression birth dates that differ by about eight years, the Oakland and Berkeley study members experienced a distinctive sequence of economic conditions and stresses. Members of the Oakland cohort were children during the relatively prosperous 1920s and entered their adolescent years as the economy collapsed. They left high school just prior to war mobilization and the beginning of World War II. Members of the Berkeley cohort experienced the Great Depression during their early years of childhood. In many cases, hard times and family insecurity did not end until their adolescence in World War II. With annual records across the depression years, data from these archival resources provide a unique opportunity to examine the short- and long-term implications of economic stress in lives. Linking Economic Stress to Children’s Lives

As shown in Fig. 1, families respond to economic loss by restructuring resources and relationships. Changing the family economy represents one effort to eliminate or modify the problem of economic hardship. Altering family rela-

Elder and Caspi

32

tionships is another way of controlling the effects of economic loss or managing its psychological consequences. Of course, some of these alterations are adaptive coping strategies, while others may release pressure without improving the situation. As we shall see below, both responses represent important linkages between economic deprivation and children’s development in the Great Depression.

Adaptations in the Family Economy Income loss altered the household economy in ways that made a profound difference in the lives of children. Drastic loss of income shifted the household economy toward more labor-intensive operations. These new modes of economic maintenance included the entry of mother and children into productive roles as earners; the involvement of children in household operations, from food preparation to laundry and cleaning; the residential doubling up of family units, and a reduction in expenditures. As a whole, these changes entailed a major shift in responsibilities from father to mother and older children. Changes in the household economy acquired unique significance in the lives of the Oakland adolescents. The older Oakland children became more valuable as conditions worsened. They were called on to meet the increased labor and economic needs of deprived households, and a large number managed tasks in the family or earned money on paid jobs. In most cases, a portion of this money went to family needs (Elder, 1974). The adolescents who obtained paid jobs were judged more responsible in financial matters by their mothers when compared to nonworking adolescents. Such work in the 1930s included “odd jobs” in the adult world, from running errands and clerking to waiting on tables, but it carried the significant implication that other people counted on them. Observers considered the employed to be more industrious and energetic than the nonemployed. They were also described as more efficacious than the nonworkers. A mother of one of these adolescent workers described him as having “one driving interest after another, usually a practical one” (Elder, 1974, p. 145). With additional responsibilities in the household, these adolescents experienced the obligations of adult status. Indeed, to observers who knew them well, they appeared to be more adult oriented in values, interests, and activities when compared to other youth. This accelerated pathway to adult status was different for boys and girls, according to their predominant role in the family economy of deprived households (Elder, Van Nguyan, & Caspi, 1985). Work roles were more common among boys and tended to enhance their freedom from traditional constraints of parental control. Moreover, the work pattern among boys extended into the adult years through an early crystallization of their work life and less floundering from one job to another. Indeed, adolescents with paid employment during the depression years were most likely to favor industry as a quality in their children at midlife when compared to other men.

Economic Stress in Lives

33

Adolescent girls were primarily involved in household responsibilities, a social adaptation that reinforced family ties and dependency while orienting them to domestic interests and roles in the adult world. As we move to the adult years of the Oakland adolescents, this household change acquired significance as an explanation of why women from hard-pressed families tended to value homemaking, family activity, and responsibility for parenting. When compared to other Oakland females, the household career of economically deprived women included relatively early family events (marriage, children), especially among those of middle-class origins, and a preference for family values at midlife. By moving adultlike responsibilities earlier, toward the adolescent years, the Great Depression accelerated the passage to adult roles. The same acceleration has been noted among contemporary children in one-parent households and in paid jobs. As Weiss (1979) has put it, children in these households grow up faster. The work experience of teenage Americans in the 1980s may have adverse consequences, from premature affluence to less investment in schooling (Greenberger & Steinberg, 1986), but these effects do not appear in our data on the Oakland cohort. A plausible reason for this difference involves the effect of hard times on families; the shift of households from capital- to labor-intensive made room for the valued contributions of children. These children had productive roles to perfom. But in a more general sense, they were needed, and, in being needed, they had the chance and responsibility to make a real contribution to the welfare of others. Being needed gives rise to a sense of belonging and place, of being committed to something larger than the self. However onerous the task may be, there is gratification and even personal growth to be gained in being challenged by a real undertaking if it is not excessive or exploitative. (Elder, 1974, p. 291).

Deprivational conditions among Oakland adolescents favored an accelerated pace toward adult status, a pace responsive to the “downward extension of adultlike obligations” in hard times, and also produced families that resembled an understaffed environment-that is, one with an excess of tasks relative to able members. Here, young and old have multiple opportunities for work and responsibility. In a series of field studies, Barker (1968) found the inhabitants of understaffed settings to be involved in more challenging and consequential actions when compared to the occupants of overstaffed environments; they are “busier, more vigorous, more versatile, and more oriented vis-a-vis the settings they inhabit, and more interdependent” (p. 190). These observations apply to adults and older children, but the picture is different for younger children. Whether the family is headed by two parents under extreme economic pressure, or by a single parent with a small income (Weiss, 1976), the effect of an understaffed family environment is apt to be undersocialization of younger children. This contrasting effect of an understaffed, economically deprived family became evident in research on the younger Berkeley children, especially males, who experienced the.economic crisis when they were more dependent on family

34

EIder and Caspi

nurturance and more vulnerable to family instability. Our findings show that the Berkeley boys who grew up in deprived households were less likely to be hopeful, self-directed, and confident about their future than youth who were spared such hardship (Elder, 1979). This dysphoric outlook is one element of a behavioral syndrome that emerged from personality ratings-personal and social inadequacy, feelings of victimization, and self-defeating behavior. In addition, Berkeley boys from deprived families held lower aspirations than children from nondeprived families, and their scholastic performance in adolescence fell well below that of adolescents from nondeprived homes. The disadvantage of growing up deprived continued in adulthood, limiting formal education among the sons of both middleand working-class parents. Whereas deprivation among the older Oakland boys led to greater mobilization of effort and ambition for adult work and family security, the same conditions among the younger Berkeley boys lowered expectations and achievements. The age-linked risk of economically stressful times for boys’ development corresponds with a similar risk among boys who experience marital discord and divorce in their family (Eme, 1979; Rutter, 1970, 1982). Similarly, in a series of longitudinal studies, Hetherington (1988) has consistently found more adverse consequences of divorce among young boys than among girls of this age. We find the very same sex difference in relation to economic stress among girls and boys in the Berkeley cohort. Various explanations have been offered for the consistent finding that, during the first decade of life, boys are more vulnerable to environmental insult. In general, males are more vulnerable to a wide range of physical hazards, and it is possible that there is a parallel biologically determined susceptibility to psychosocial stressors (Earls, 1987). It is also possible that boys’ vulnerability to family stress may be a consequence of their greater exposure to discord (Hetherington, Cox, & Cox, 1978) or of parents’ different responses to problem behavior in boys than in girls (Maccoby & Jacklin, 1983; Snow, Jacklin, & Maccoby, 1983). Boys are more likely to react to family stress with disruptive behaviors, a type of reaction that is more likely to elicit a negative response from parents (Emery, 1982). Moreover, assuming that control over the environment is more important for young boys than for girls, the environment of a discordant family in stressful times may be more disturbing for boys (Gunnar-Vongnechten, 1978). Block, Block, and Morrison (1981) have suggested, in addition, that the salience of the two parents may differ for boys and girls; for girls, the lesser salience of the father may attenuate the effects of family discord. This seems especially plausible in light of household changes in the Great Depression, since fathers’ loss of earnings and resulting adaptations in family support increased the relative power of mothers concerning affection, authority, and completion of basic tasks, and diminished the attractiveness of fathers as role models.

Economic Stress in Lives

35

At this point, we can draw no firm conclusions about the precise reasons for the heightened vulnerability of young boys to economic stress in the family, and the true explanation probably reflects complex interactions among the processes enumerated here. Such findings, however, pose important questions for further research, especially about changes in family relationships during stressful times and their effects on younger children. Alterations in Family Relationships Economic setbacks clearly make a difference in psychological functioning (for reviews see Dooley & Catalano, 1980, and Horwitz, 1984). But how does income loss influence children’s behavior and development? To answer this question, we turn to family relationships. The adverse consequences of stressful economic times do not necessarily have direct effects. They are more often produced indirectly through their disorganizing effects on family relationships. The schematic model shown in Fig. 2, based on findings from a series of studies, provides some insight into the dynamics of family relations under stress and the increased vulnerability of children in stressful times. Our model linking stressful events to family dynamics includes personality characteristics, qualities of relationships, and reciprocal influences between parents and their children. Personality is defined here as a characteristic way of responding to environmental stimuli, a relatively enduring quality of individuals. Like personality, relationships have their own developmental history. Although partly dependent on the unique dispositions of individuals engaged in interaction, relationships evolve over time as emergent interaction patterns that cannot be explained solely in terms of individual personalities (Hinde, 1979).

A+ Economic

Marital Conflicts

5 7

Fig. 2. The dynamics of family behavior under stress.

36

Elder and Caspi

More specifically, we have focused on three qualities of disorganized families in stressful times: personal instqbility, marital tension, and arbitrary, inconsistent parenting. These variables are assumed to mediate the effects of extrafamilial stressors on the child. To the extent that economic stress increases unstable behavior, marital discord, and rates and levels of unpredictable, hostile exchanges, children are at risk. On the other hand, as long as unstable behavior is kept in check, as long as marital conflicts are resolved as usual, and as long as discipline remains consistently applied, children should be protected from the psychological costs of environmental stressors. On a general level, our analytic model bears some resemblance to Patterson’s (1982, 1988) schematic outline of factors contributing to antisocial child behavior: a major crisis, such as unemployment, increases the likelihood of a disruption of family management practices, which in turn increases the risk of antisocial child behavior. The key, of c o m e , is to disentangle the various elements in families characterized by multiple problems. Because aversive interactions, whether between spouses or between parents and children, so frequently occur in the context of a disordered family, their effects cannot be easily separated from the total pattern. We shall make an effort here to summarize what we do know about the process of influence. Changes in Marital Relations

The most visible change in family relations is separation or divorce, and studies c o n f i i the relationship between unemployment, low income, and marital dissolution. For example, it is estimated that separation rates are twice as high among families where the husband is unemployed as in those experiencing stable employment (Ross & Sawhill, 1975). But even where economic stress does not lead to separation or divorce, marital discord and conflict may still increase (Moen, Kain, & Elder, 1983). What remains unclear is the process by which economic stress makes a difference in marital relations: What part of the influence of economic loss on families occurs directly through disrupted patterns of social interaction, and what part occurs through adverse psychological changes in one or more members? If interpersonal relations become strained, is this a source of individual distress? How do individual adaptations enter this sequence? Studies of the Berkeley couples and their families indicate that income loss generally weakened marital ties by producing adverse change in the husbands’ behavior. In addition, men who were unstable before the Depression frequently became more irritable and explosive following a heavy loss of income, and these changes further increased marital tension and conflict (Elder, Liker, & Jaworski, 1984; Liker & Elder, 1983). To explicate this process, we begin with zero-order correlations between

Economic Stress in Lives

37

economic deprivation and measures of marriage interactions and personality from 1930 to 1941. These measures were obtained from staff assessments of parents based on interviews with the Berkeley mothers. Marital tension was indexed by a 5-point scale, with values that indicated a relationship characterized byfrankness, aflection, and agreement at the low end, and by disruptive conflicts at the other extreme. Financial conflicts were indexed by a 5-point scale with values that indicated perfect agreement, no friction at the low end, and acute disagreement and open conflict at the high end. Instability was indexed by two 7-point ratings: irritability and nervous stability. Irritability ranged from extremely even tempered (low) to explosive; nervous stability ranged from exceptionully even keel, even in the face of trying circumstances to the presence of behaviors interfering with social functioning. As shown in Table 1 (lines A and B), the economic correlates of marital tension and financial conflicts were modest in 1930. As expected, income loss correlated more strongly with the marriage indicators during the economic trough later in the 1930s. And not surprisingly, economic deprivation was more strongly related to financial conflicts than to the global assessment of marital tension. Correlations between economic deprivation and husbands’ personality (line C of Table 1) are presented separately for subgroups of husbands low or high in personal instability in 1930. This procedure enables us to examine the accentuation hypothesis-that there is an increase in emphasis of an already prominent characteristic due to its reinforcement by selected contexts and events (Feldman & Weiler, 1976). As outlined in Fig. 1, we expected economic hardship to produce more inadequacy among the unstable families and possibly even greater competence among the stable ones. Overall, the correlations by category support the view that unstable behav-

Table 1. Correlations at Four Points in Time Between Marriage and Personalitv Measures and Income Loss0 ~

Correlations with uercentaee income loss. 1929-1933 Measure ~

~~

A.

Marital tension

B.

Financial conflicts

1930 .13 (109)

.20* (108)

C.

Husband instability Below median (1930) Above median (1930)

-.I7 (57) .02 (51)

ONs are shown in parentheses, *p

< .05.

1933- 1935

1936- 1938

~~

.19* (97) .36* (94)

- .08 (53) .42* (47)

1939- 1941

38

Elder and Caspi

ior in the 1930s depended on the joint influence of economic deprivation and psychological resources. Men who entered the 1930scalm and stable did not as a rule become explosive and moody under intense economic pressure. Indeed, for this group the correlations were generally negative, suggesting that men who were successful copers moderated any inclination toward erratic behavior when they lost jobs and income. The opposite group became increasingly unstable under economic deprivation: Men who were irritable, tense, and moody before hard times displayed an accentuation of these characteristics. In sum, the correlations document a pattern of interaction effects: The effects of economic deprivation depended on the interaction of psychological resources and income loss. A more rigorous test of this interaction by regression analysis is presented below. To investigate this conditional hypothesis about change in individual behavior and marital relations, we entered an interaction term-income loss by 1930 instability-into two regression models, one predicting 1933- 1935 instability and the other predicting 1933- 1935 marital conflicts. The interaction effects provided support for the accentuation hypothesis. Personal instability before income loss made a reliable difference in the effect of income loss on the behavior of men during the depression (t = 1.77, p < .lo). Hard times notably accentuated the unstable tendency of husbands who wen? relatively unstable and imtable before the Great Depression. In addition, husbands’ instability before income loss interacted with income loss in affecting marital tension during the 1930s ( t = 2.01, p C .05). Income loss sharply increased marital tensions, and it did so primarily in families where men were unstable at the outset of the crisis. Although these findings do not relate the psychological distress of parents to their relations with children, they represent an important building block for understanding how unstable fathers and discordant marriages, in combination with punitive parenting, mediate the effects of economic stress on children’s social development. Changes in Parent-Child Relations

Some implications of economic change for parent-child relations have been noted already in our discussion of change in the family economy. Other research suggests, in addition, that economic factors are significantly related to child maltreatment (Kadushin & Martin, 1981; Steinberg, Catalano, & Dooley, 1981; Straus, Gelles, & Steinmetz, 1980). The meaning of these correlates, however, remains ambiguous despite repeated documentation. For example, it is not clear whether the behavior of abusive parents causes or is caused by work-life instability. As we noted in the introduction, this ambiguity is merely one facet of a larger deficiency in research on economic change: that of underspecified or incompletely developed models. Using data on children’s problem behavior from the Berkeley archive, we

Ecowmic Stress in Lives

39

identified two behavior patterns that are likely to be influenced by aversive interactions in the family, and in turn, to shape the course of children’s social relationships (Elder, Caspi, & Downey, 1986; Elder, Liker, & Cross, 1984): 1. The dzficulr child syndrome refers to the intense, high-strung child whom parents find hard to manage. Such temperamental difficulties are often accentuated by parental intolerance, inconsistency, and conflict over child rearing (e.g., Thomas & Chess, 1977). In the Berkeley study this syndrome was indexed by the average of three 5-point ratings obtained from clinical interviews with mothers. They rated the extent to which their child was quarrelsome, having “a chip on the shoulder” and instigating quarrels with no apparent provocation; negativistic, displaying a compulsive and habitual urge to do the opposite of what is expected; and irritable, i.e., explosive and overreactive. 2. The temper tantrum syndrome was based on two other ratings by the mother: severity and frequency of temper tantrums. Severe tantrums involved biting, kicking, striking, throwing things, and screaming-with anger completely dominating behavior. Tantrum frequency ranged from one per month to several times a day. The temper tantrum syndrome measure captures a key identifying characteristic of children classified as socially aggressive (Achenbach, 1978). Moreover, severe and frequent tantrums have been shown to be highly contingent on the disciplinary strategies of parents (Patterson, 1982). With panel data over an eight-year period (1930-1938), our research has shown that the influence of family income loss on the problem behavior of children is indirect. Drastic income loss increased the prospect of children’s problem behavior, but only by increasing aversive interactions within the family. Indeed, the master link between environmental stress and children’s problem behavior was inconsistent and arbitrary parenting. Arbitrary behavior (rated by the Institute staff on the basis of maternal interviews) refers to inconsistent discipline that expresses the parental mood rather than a response to the child’s behavior; its result is that the child never knows what to expect. Inconsistent parenting has been repeatedly implicated in the development of externalizing problems (Loeber & Stouthamer-Loeber, in press), and our findings suggest that economic change may well increase the probability of arbitrary behaviors by fathers. As long as discipline is consistently applied, children seem to be highly resilient under a wide range of parental styles. It is the inconsistent and arbitrary patterns that appear most problematic for healthy development in families under stress (Rutter, 1980). Stressful background factors appear to raise the level of aversive interactions between parents and children, and to reduce parents’ ability to deal effectively with their developing children. This, however, is most likely an interactive process, and here we face the limitation of a unidirectional influence model. It is highly probable that children residing in an aversive system also make a significant contribution to the continued disruption of the family.

40

Elder and Caspi

Indeed, from research completed on the Berkeley cohort (Elder et al., 1984), it is clear that young children (about 18 months old) who were observed as displaying problematic behavior before the depression experienced a greater risk of arbitrary and punitive treatment by fathers during the worst years, in comparison to less problematic children. Problem children in 1930 were not more likely to have fathers with hostile feelings at that time, but three years later in the depression they were significantly more likely to receive arbitrary discipline by fathers than were nonproblematic children. These correlations remained significant even with adjustments for the precrisis instability of fathers. As we have seen, economic stress was often a change agent of family life, and fathers’ behavior emerged as a critical link between depression hardship and the family experience of children. Surprisingly absent in this account was mothers’ behavior. Of course, the Berkeley children entered the depression crisis with various relationships to their mothers. Some were recipients of maternal warmth and affection, while others lacked such nurturance. The likely implications of this predepression variation are suggested by the dependency of young preschool children on their family environment and especially on the mother. Maternal responsiveness to child cues affects both immediate and long-term adaptations by children (Schaffer, 1977), and in times of stress, the emotional support of mothers can protect children from conflict and overwhelming demands (Caplan, 1976). Using data from the Berkeley archive, we have examined the effects of family hardship and punitive fathers on the behavior of children whose mothers differed on expressed affection toward their children (Elder, Caspi, & Van Nguyen, 1986). To investigate the moderating influence of this maternal sentiment, we stratified the Berkeley sample by a predepression (1930) measure of mother’s demonstration of affection toward the child. For purposes of analysis, we classified mothers as “undemonstrative” if they had scores below the median and as “affectionate” if they had scores above the median. The mothers’ precrisis attitude toward their children was found to have significant implications for fathers’ behavior under conditions of severe income loss. The results in Fig. 3 indicate that we gain valuable information about the determinants of arbitrary parenting by treating these contrasting family subgroups separately. In particular, a family context defined by an initially undemonstrative mother was more likely to lead to fathers’ arbitrary discipline, especially when prompted by income loss. Moreover, income loss was positively related to the temper outbursts of children only in families with undemonstrative mothers. Thus mothers’ expressive affection vs. undemonstrativeness, measured prior to the depression, tells us a good deal about the influence of economic hardship on child disturbance and on the nature of fathers’ actions under stress. The overall pattern suggests a lack of regulation in the household or the absence

Economic Stress in Lives A.

41

Mother affectionate. 1930 (N.49)

B. Mother undemonstrative,

1930 (N=44)

Fig. 3. Linking family hardship to children’s lives by maternal demonstrativeness. (Note: Regression coefficients in standardized form. Zero-order correlations in parentheses. + p < .lo; * p < .05; **p

< .01.)

of effective normative constraints in hard times. All paths leading to child disturbance were stronger in the group of children with undemonstrative mothers, and the total variance explained was three times that of the corresponding model for affectionate mothers (R2= .30 vs. .09). Indeed, the aversive dynamics of family stress were relatively limited when mothers were affectionate toward the study child. Family units with strong affective mother-child bonds restrained the arbitrary treatment of children by depression fathers, and children with warm, affectionate mothers were under some protection from economic hardship. Mothers’ initial attitude toward the study child thus differentiated contrasting family trajectories under the economic pressures of the 1930s. One path led toward the maltreatment of children and their subsequent problem behavior; the other traced a more benign course in which the child was sheltered from the adverse influence of depression hardship and the punitive behavior of fathers. The general patterns noted here have been observed in other recent studies of children at risk. For example, Rutter (1985)has noted that the presence of one good parent-child relationship serves to reduce the risk of conduct disorders associated with family discord. The precise mechanisms associated with this

Elder and Caspi

42

protective factor are difficult to disentangle. Our results suggest one possible process, in which the parent with whom the child had a good relationship actively sought to ensure that the child was out of the way of major discord. Another possibility, however, is that the security of one good relationship increased the child’s self-esteem and that this exerted the protective effect (Rutter, 1985). Whatever the underlying mechanisms, the protective effects of relationships within the family clearly deserve greater attention in studies of economic stress.

Summary and Implications for Research Stressful economic times bring two issues to mind the process by which economic hardship entails substantial developmental risk for children, and the factors that minimize or accentuate such risk. Our goals have been to specify linkages between macroeconomic change and children’s development by tracing deprivational effects through family adaptations, and to examine empirically these mediating processes as well as the factors that moderate them, Accordingly, the presentation of results from our studies displays progressively more complex elements in a model of the process by which families and children adapt to stressful economic conditions. We have argued that social and macroeconomic change is linked to children’s lives through alterations in the family. In particular, we have identified two micromodels of how a drastic economic loss influences children’s development. The first model involves the consequences for family relationships of a shift in the family economy from capital- to labor-intensive operations. The second model involves the consequences of alterations in family relationships, both spousal and parent-child, for children’s social and personality development. In addition, we have argued that these changes take place in a reciprocal social system that involves mutual accommodation and adjustment between family members. Thus, changes in any relationship or in any one family member’s behavior inevitably influence other relationships and other family members. For illustrative purposes, we have addressed here only some of the relevant factors (e.g., father’s instability, the quality and nature of the mother-child relationship); other factors are undoubtedly important. These illustrations, however, demonstrate a more general point. It is that the social costs of economic stress can be understood only in relation to the reso’urcesand attributes-psychological and social, individual and relational-that are brought to change situations, and that therefore the influence of environmental stressors on individual functioning must be considered in terms of the family context in which they operate. Whether research efforts are oriented toward understanding the longterm effects of historical change in the 1930s or the short-term consequences of plant closings in the 1980s, the analytic principles we have outlined remain central to the ecological study of human development. They define research

Economic Stress in Lives

43

priorities for better understanding the social costs of economic stress for children.

References Achenbach, T. M. (1978). The child behavior profile, I: Boys aged 6 through 11. Jownaf of Consulting and Clinical Psychology. 46, 478-488. Bane, M. J., & Ellwood, D. T. (1986). Slipping into and out of poverty: The dynamics of spells. Journal of Human Resources, 21, 1-23. Barker, R. (1968). Ecological psychology. Stanford, CA: Stanford University Press. Block, J. H., Block, J., & Momson, A. (1981). Parental agreement-disagreement on child-rearing orientations and gender-related personality correlates in children. Child Development, 52, 965-974.

Campbell, A. (1981). The sense of well-being in America: Recent patterns and trends. New York: McGraw-Hill. Caplan, G. (1976). The family as support system. In G. Caplan & M. Killiken (Eds.), Support systems and mutuul help: Multidisciplinary explorations. New York: Grune & Stratton. Dooley, D., & Catalano, R. C. (1980). Economic change as a cause of behavioral disorder. Psychological Bulletin, 87, 450-468. Duncan, G. J. (1984). Years ofpoverty, years qfplemy. Ann Arbor, MI: Institute for Social Research. Duncan, G. J. (1987). The volatility of family income over the life course. In P. B. Baltes, D. L. Featherman, & R. M. Lerner (Eds.), Life-spandevelopment and behvior (Vol. 7). Hillsdale, NJ: Erlbaum. Duncan, G. J., & Morgan, J. (1985). The panel study of income dynamics. In G. H. Elder, Jr. (Ed.), Life course dynamics: Trajecrories and transitions (pp. 50-71). Ithaca, NY: Cornell University Press. Earls, F. (1987). Sex differences in psychiatric disorders: Origins and developmental influence. Psychiatk Developments, I, 1-23. Eichorn, D. H. (1981). Samples and procedures. In D. H. Eichom, J. A. Clausen, N. Haan, M. P. Honzik, & P. Mussen (Eds.), Present andpast in middle life. New York: Academic Press. Elder, G. H.,Jr. (1973). On linking social structure and personality. American Behavioral Scientist, 16, 785-800.

Elder, G. H., Jr. (1974). Children of the Great Depression. Chicago: University of Chicago Press. Elder, G. H., Jr. (1978). Approaches to social change and the family. American Journal of Sociology, 84, 1-38. Elder, G. H., Jr. (1979). Historical change in life patterns and prknality. In P.B. Baltes & 0.G. Brim, Jr. (Eds.), Life-span development and behavior (Vol. 2, pp. 117-159). New York Academic Press. Elder, 0 . H.,Jr. (1981). Social history and life experience. In D. P. Eichorn, J. A. Clausen, N. Haan, M. P. Honzik, & P. Mussen (Eds.), Present andpasr in middle life (pp. 3-31). New York: Academic Press. Elder, G. H., Jr., & Caspi, A. (in press). Human development and social change: An emerging perspective on the life course. In N. Bolger, A. &pi, G. Downey, & M. Moorehouse (Eds.), Persons in context: Developmental processes. New York: Cambridge University

Press. Elder, G. H., Jr., Caspi, A., 8t Burton, L. M. (1988). Adolescent transitions in developmental perspective: Sociological and historical insights, In M. Gunnar (Ed.), Minnesota symposia on child psychology (Vol. 21). Hillsdale, NJ: Erlbaum. Elder, G. H., Jr., Caspi, A., & Downey. G. (1986). Problem behavior and family relationships: Life course and intergenerational themes. In A. Sorensen. F. Weinert, & L. S h e d (Eds.), Human development and the lije course: Multidisciplinaryperspectives (pp. 293-340). Hillsdale, N J Erlbaum. Elder, G. H., Jr., Caspi, A., &Van Nguyen, T. (1986). Resourceful and vulnerable children: Family influence in stressful times. In R. K. Silbereisen, K. Eyferth, & G. Rudinger (Eds.), Devel-

44

Elder and Caspi

opment as action in context: Problem behavior and normal youth development. New York: Springer. Elder, G. H., Jr., Liker, J. K., &Cross, C. E. (1984). Parent-childbehavior in theGreat Depression: Life-course and intergenerationalthemes. In P. B. Bakes & 0. G. Brim, Jr. (Eds.), Life-span development and behavior (Vol. 6 , pp. 109-158). New York Academic Press. Elder, G. H., Jr., Liker. J. K., & Jaworski, B. J . (1984). Hardship in lives: Historical influences from the 1930s to old age in postwar America. In K. McCluskey & H. Reese (Eds.). Lifespan developmental psychology: Historical and cohort effects (pp. 161-201). New York Academic Press. Elder, G. H., Jr., Van Nguyen, T., & Caspi, A. (1985). Linking family hardship to children’s lives. Child Development, 56, 361-375. Eme, R. F. (1979). Sex differences in childhood psychopathology: A review. Psychological Bulletin, 86, 574-595. Emery, R. E. (1982). Interparental conflict and the children of discord and divorce. Psychological Bulletin. 92. 310-330. Feldman, K. A,, & Weiler, J. (1976). Changes in initial differences among major-field groups: An exploration of the “accentuation effect.” In W. H. Sewell, R. M. Hauser, & D. L. Featherman (Eds.), Schooling and achievement in American sock@ (pp. 373-407). New York Academic Press. Furstenberg, F. F., Jr., Brooks-Gunn, J., & Morgan, P. S. (1987). Adolescent mothers in later life. New York Cambridge University Press. Greenberger, E., & Steinberg, L. (1986). When teenagers work: The psychological and social costs of adolescent employment. New York: Basic Books. Gum-Vognechten. M. (1978). Changing a frightening toy into a pleasant toy by allowing the infant to control its actions. Developmental Psychology, 14, 157-162. Hetherington, E. M. (1988). Parents, children and siblings six years after divorce. In R. A. Hinde & J. Stephenson-Hinde(Eds.), Relationships withinfamilies (pp. 31 1-331). New York Oxford University Press. Hetherington, E. M., Cox, M., & Cox, R. (1982). Effects of divorce on parents and children. In M. Lamb (Ed.), Nonfraditional families (pp. 233-288). Hillsdale, NJ: Erlbaum. Hinde, R. A. (1979). Towards understanding relationships. London: Academic Press. Horwitz, A. (1984). The economy and social pathology. Annual Review of Sociology, 10, 95-119. House, I. S.(1981). Social structure and personality. In M. Rosenberg & R. H. Turner (Eds.), Social psychology: Sociological perspectives (pp. 525-561). New York: Basic Books. Kadushin, A. K., & Martin, J. A. (1981). Chi1dabuse:An interactionalevent. New York: Columbia University F‘ress. Kelvin, P., & Jarrett, J. E. (1985). Unemployment: Its social psychological effects. London: Cambridge University Press. Keyssar, A. (1985). Out of work. Cambridge, MA: Harvard University Ress. Kohn, M., & Schooler, C. (1983). Work and personality: An inquiry into the impact of social stratification. Norwood, NJ: Ablex. k b e r , R., & Stouthamer-Loeber, M. (in press). Family factors as correlates and predictors of juvenile conduct problems and delinquency. In M. Tonry & N. Moms (Eds.), Crime and justice (Vol. 7). Chicago: University of Chicago Press. Liier, I. K., 8r Elder, G. H., Jr. (1983). Economic hardship and marital relations. American Sociological Review, 48, 343-359. Maccoby, E. E., & Jacklin, C. N. (1983). The “person” characteristicsof children and the family as environment. In D. Magnusson & V. L. Allen (Eds.),Human development: An interactional perspective (pp. 75-91). New York: Academic Press. Minuchin, P. (1985). Families and individual development: Provocations from the field of family therapy. Child Development, 56, 289-302. Moen, P., Kain, E., & Elder, G. H., Jr. (1983). Economic conditions and family life: Contemporary and historical perspectives. In R. Nelson & F. Skidmore (Eds.), American families and the economy (pp. 213-259). Washington, DC: National Academy Press. Oppenheimer, V. K. (1982). Work andfamily: A study in social demography. New York Academic Press.

Economic Stress in Lives

45

Patterson, G. R. (1982). Coercive family process. Eugene, OR: Castalia. Patterson, G . R. (1988). Family process: Loops, levels, and linkages. In N. Bolger, A. Caspi, G. Downey, & M. Moorehouse (Eds.), Persons in conrexr: Developmental processes (pp. 114-151). New York: Cambridge University Press. Ross, H. L., & Sawhill, L. Y. (1975). Time of rransition: The growth offamifies headed by women. Washington, DC: Urban Institute. Rutter, M. (1970). Sex differences in children’s responses to family stress. In E. I. Anthony & C. Koupernik (Eds.), The child in his family (Vol. 1, pp. 165-196). New York: Wiley. Rutter, M. (1980). Changing youth in a changing society. Cambridge, MA: Harvard University Press. Rutter, M. (1982). Epidemiological-longitudinal approaches to the study of development. In W. A. Collins (Ed.), Minnesota symposia on child psychology (Vol. 15, pp. 105-144). Hillsdale, NJ: Erlbaum. Rutter, M. (1985). Resistance in the face of adversity: Protective factors and resistance to psychiatric disorder. Brirish Journal of Psychiatry. 147, 598-61 1 . Schaffer, (1977). Mothering. Cambridge, MA: Harvard University Press. Snow, M. E., Jacklin, C. N., & Maccoby, E. E. (1983). Sex-of-child differences in father-child interaction at one year of age. Child Developmenr. 54, 227-232. Steinberg, L. D., Catalano, R., & Dooley, D. (1981). Economic antecedents of child abuse and neglect. Child Development, 52, 975-985. Straws, M. A., Gelles, R. J., & Steinmetz, S. K. (1980). Behindclosed doors. New York: Anchor. Thomas, A,, & Chess, S. (1977). Temperament and developmenr. New York: Brunner/Mazel. Thomas, W. I. (1909). Sourcebook for social origins. Boston: Badger. Wachs, T. (1979). Proximal experience and early cognitive-intellectual development: The physical environment. Merrill-Palmer Quarrerly, 25, 3-41. Weiss, R. (1976). Marital separation. New York: Basic Books. Weiss, R. (1979). Growing up a little faster: The experience of growing up in a single-parent household. Journal of Social Issues, 35(4), 97-1 1 1 . Weitzman, L. (1985). The divorce revolution. New York: Basic Books.

GLEN H. ELDER, JR., is Howard W. Odum Distinguished Professor of Sociology and Research Professor of Psychology at the University of North Carolina, Chapel Hill. After completing his doctorate at Chapel Hill, he served on the faculties of the University of California at Berkeley and Cornell University. With a Senior Scientist Award from NIMH, he is continuing his longitudinal research on social change in the life course and mental health. AVSHALOM CASPI is Assistant Professor of Psychology at Harvard University. He received his Ph.D. in developmental psychology from Cornell University. He is currently involved in several longitudinal studies investigating the causes and consequences of continuity in problem behavior across the life course.