Gary Becker is arguably one of the most influential academics working in the area of the family. Recognition of his work among both economists and noneconomists culminated in Becker being awarded the 1992 Nobel Prize in Economics. His seminal work on the family covers topics such as marriage formation and dissolution, fertility, investment in children, education, intergenerational mobility, allocation of time and resources in the household, and the role of altruism. Economic concepts emphasized by Becker such as the opportunity cost of time have become common parlance among family researchers and policy makers. In this article, I outline the economic approach to family behavior pioneered by Becker, [1] assess the criticisms of this approach made by both economists and noneconomists, evaluate the contributions of this approach to our understanding of family processes and child outcomes, and explore its potential in fostering interdisciplinary work on these topics.
I focus broadly on two areas: household production and altruism. The first area contributes to an understanding of the activities the family undertakes and how these activities are accomplished. The household Production approach has been widely used to study educational outcomes, health production, recreation demand, and the cost of children. The second area explicitly recognizes that a family is made up of different individuals. Altruism is used to explain why the family exists as an institution and how family goods and services are allocated to the individuals in the family in the absence of market or legal mechanisms that determine these allocations in other settings.
HOUSEHOLD PRODUCTION
The economic approach to the study of human behavior posits that individuals have certain goals and preferences and that they try to do the best they can, given the constraints that they face. [2] These constraints include time, income, prices, technology, information, government laws, and social norms. The first law of economics, the law of demand, states that when the cost of an activity increases, the demand or desire to undertake that activity will decrease. This law comes from the idea that there are competing desires, but limited resources. When the cost of one activity increases, individuals may be better off trying to fulfill their basic desires in an alternative way. In addition, economists argue that an increase in income (holding all else constant) will generally increase the demand for most kinds of activities or commodities. Economists are often criticized for the assumptions of rationality and utility maximization, but these two assumptions imply little more than what is stated in general terms above. Mathematical models are used to formalize these assumptions, explicitly derive the nature of the constraints, and make additional assumptions that allow additional insights, but many economic insights can be obtained solely from the two basic economic propositions--the law of demand and positive income effects.
The tricky part in any analysis is in specifying and measuring the costs, benefits, and constraints. The household production approach provides a framework for doing this. The approach describes the family as a small factory that produces commodities of value to the family. [3] These commodities are defined very generally and include such things as children, health, and leisure. Commodities are produced with inputs of family members' time and money or goods purchased in the market. The production process also depends on the available technology and environmental factors such as the neighborhood in which a family lives.
As the name implies, the primary focus of this model is the production of commodities, rather than the distribution of those commodities to various family members. The model therefore ignores the possibility that individual family members may have different preferences and may behave strategically to maximize their individual well-being at the expense of others. Instead, the model assumes that family members have common preferences or that there is only one decision maker in the family. Thus family income/consumption possibilities will be maximized, and this action benefits all family members.[4]
There are four essential components to a household production model: the utility function, the production function, the time constraint, and the money constraint. The utility function is a standard tool used in economic analysis. It specifies how a family values various commodities and what kind of trade-offs a family would be willing to make among the various commodities that it values. The production function also has a long history of use in economic analysis. The innovation in Becker's work is in the kinds of commodities that are produced. Instead of specifying a production function for automobiles or corn, the commodities analyzed in household production models are more intangible, such as home-cooked meals, love, a solid marriage, well-adjusted and healthy children, and so on. Because these commodities are not sold in the marketplace, we cannot directly observe their price. We refer to the cost of producing a particular commodity as its shadow price. The time and money constraints reflect the fact that these resources are limited. Time or money used in the production of one commodity are no longer available to be used in the production of another commodity.
In this framework, families have two basic kinds of decisions to make:
1. How to produce each commodity (i.e. what inputs and technology to use)
2. How much of each commodity to consume.
Regarding the production decision, the model predicts that families will use the technology that is the most efficient for a given outcome (i.e. has the lowest cost or shadow price). For a given technology, the shadow price of a commodity will depend on the price of inputs (both time and money) and on the productivity of those inputs. The consumption decision will depend both on the costs of producing various commodities and on tastes for one commodity relative to others.
To be more specific, let's take children as an example. A count of the number of parenting books on the library shelves provides some indication of the large variety of approaches to child rearing. The specific child-rearing (production) method used differs across families and cultures and over time. How does a family decide which child rearing method to use?
To answer this question, compare two very different methods of child rearing. In the first method, the mother stays home full-time to care for the child.[5] Monetary inputs include such things as clothing, food, shelter, and toys. We characterize this method as time-intensive (in terms of the mother's time). In the second method, the mother works full-time in the labor market and spends some time with the child after work. Monetary inputs for this method include purchased child care as well as clothing, food, and so on. We characterize this method as goods or money intensive.
The economic approach predicts that the family chooses the technology that yields the lowest cost. The cost is calculated as the amount of mother's time used per child multiplied by the cost of that time and the amount of money or market goods multiplied by the cost of those goods. The cost of time is usually measured by the wage rate, the amount of money the mother would have to forgo to spend time at home with the child. The cost of child rearing then depends on the cost of each input and amount of inputs needed to produced a given output (i.e., the productivity of the inputs). Everything else held equal, if a woman could earn per hour working in the market, and if the cost of full-time child care were .50 per hour, she would stay at home and choose the time-intensive method of child rearing.[6] Alternatively, if she could earn per hour, she would enter the labor market and choose the goods-intensive method of child rearing. A number of recent articles have shown that mothers who face higher child care costs are less likely to enter the labor force and work fewer hours when they are in the market (Averett, Peters, & Waldman, 1995; Blau & Robins, 1988; Connelly, 1992).
Of course, some mothers with higher wages will still choose to stay home. This could happen for several reasons. First, it is possible that purchased market child care is not of comparable quality to the child care that some mothers could provide. To put it another way, in some cases, purchasing the same quality of child care as provided by the mother would be extremely expensive. Families differ in the quality of child care that can be provided at home. For example, more experienced mothers may be able to perform child-rearing tasks with less effort, time, and worry, because they have already learned the requisite skills. Similarly, more educated mothers may be more productive in child rearing, because they are able to process information and learn new skills more quickly. There is some evidence in the child development literature that high-quality child care programs can compensate for the poor home environment of children from disadvantaged (i.e., low income, low parental education) families (Hayes, Palmer, & Zaslow, 1990). In contrast, it may be more difficult or expensive to find adequate substitutes for the child care provided by some nurturing and highly educated mothers. Evidence that mothers with higher home productivity (measured by maternal education) spend more time in child care and that the increased child care time of highly educated mothers leads to better child outcomes (Datcher-Loury, 1988) is consistent with the above argument.
A second reason that some higher wage mothers stay at home is because of tastes for child rearing versus work. Some women like their jobs (independent of what they get paid), whereas others like caring for children. The economic approach recognizes differences in tastes but argues that these differences are not observable by the analyst or policy maker.[7] The economic approach is, therefore, not appropriate to use in analyzing individual differences. Rather its strength lies in predicting the effects of changes in prices, income, technology, and other constraints on the average behavior of individuals with similar characteristics.
The increase in female wages over time (both in absolute terms and relative to male wages) provides a good example of a change in the price of inputs. The household production framework outlines two responses to input price changes. First, because the shadow price of the commodity has risen, the demand for that commodity--in this example, children--should fall. This effect is called substitution in consumption. Second, an increase in wages makes time-intensive methods of child rearing more expensive, so some households will switch to goods or money intensive methods (i.e., purchase child care in the market). This effect is called substitution in production. Economic theory tells us the conditions that affect the size of each response, but the substance of other disciplines allows us to fill in the details and make actual predictions.
Substitution in production will depend both on the child care market and on the production process. The degree of substitution in production will be smaller when the alternatives to maternal care are not very good substitutes. The economic choice model predicts that mothers faced with poor substitutes would be unlikely to reduce their own child care time, even if their wages increased substantially. Another factor that affects the degree of substitution in production is the supply of alternative care providers. If that supply is inelastic--that is, the quantity available is relatively fixed--substitution in production will be smaller. Until recently, nonmaternal child care was typically provided by relatives. If the mother did not have any relatives living nearby who were willing to provide care, there were few other good alternatives to her own time in child care. Over the last few decades, a child care market has developed, and there are a whole variety of substitutes for mother's time, from in-home child care to family day-care providers to day-care centers. The possibilities for substitution in production have expanded enormously. The child development literature has studied the characteristics of providers and centers that produce high-quality outcomes (see Hayes et al., 1990, for a summary of recent research).
The degree of substitution in consumption that is possible depends on the nature of the utility function. If other commodities are good substitutes for children, the demand for children will fall substantially when the cost of children increases. Demographers have suggested a number of reasons for having children. These include (a) provision of family labor, (b) old age security, (c) companionship and love, and (d) continuation of the family name. The first two reasons were more prominently cited in surveys of less developed countries and in countries with less well-developed social infrastructure (World Development Report, 1984). With modern technology, however, machines can do the farm work that children used to provide. Social security provides a good substitute for old-age financial support. If children are desired for these reasons, we would expect a large fertility response to an increase in female wages and labor market opportunities. There are arguably fewer substitutes for the latter two categories. If children are desired for those reasons--love and legacy--an increase in female wages would have only a small effect on fertility rates.
The distinction between quantity of children (i.e., fertility rates) and quality of children is also relevant in the context of substitution in consumption. By quality, economists usually mean the amount of expenditures (both money and time) on children, but the theoretical concept actually refers to education, health, social, emotional and cognitive development, and other aspects of child well-being that could result from the higher expenditures. When child mortality rates are low, the quality of children may be a good substitute for quantity of children.[8] Because very young children require constant care, but intellectual development can be stimulated by market purchases such as high-quality schooling, books, and computers (in addition to parental time), quality may be less time-intensive than quantity,[9] and an increase in wages could cause a large decrease in fertility rates.
To model the trade-offs between quantity and quality, Becker specified a family budget constraint where the amount spent on children is equal to the number of children multiplied by the average resources invested per child, an indicator of quality. If a family chooses a higher level of quality (i.e., higher average resources per child), the shadow price of a child is higher. For example, compare child-rearing costs for families who do and do not send their children to college. The cost of an additional child differs for the two families by the cost of a college education. Similarly, the shadow price of a college education is higher for larger families, because they must pay for college tuition for more children. The price of quantity (number of children) is a function of the chosen level of quality, and the price of quality is a function of the chosen level of quantity. If the demand for quality increases, the price of quantity will increase, which will, in turn, decrease the number of children demanded. This decrease in quantity will decrease the price of quality and have a further effect on the demand for quality, and so on. This quantity/quality interaction depends on the assumption that parents care equally about all of their children and want to invest an equal amount of resources in each child.[10]
The trade-offs between quantity and quality have long been noted in the sociological literature.
Blake (1989) explained the negative correlation between family size and children's educational attainment as being caused by the dilution of family resources as family size increases. Becker's model has an additional implication: Small changes in the price of quantity or quality--perhaps due to changes in female wage rates, subsidies to education, or decreases in child mortality rates---can have large effects on fertility rates and the demand for quality. To illustrate the effects of quantity/quality trade-offs, Becker presented aggregate data for several countries and time periods, showing that large decreases in fertility were associated with large increases in schooling (Becker, 1991, pp. 148, 151).
If there were good substitutes for numbers of children other than child quality, the continued increase in female wage rates would lead to continual declines in fertility rates and increases in the proportion of women who remain childless. If, instead, the substitution of quality for quantity were an important part of the story, we would expect fertility rates to decline and then reach some low asymptote, because a family has to have at least one child to have any child quality. With the exception of the baby boom, fertility rates in the United States declined dramatically over the first three quarters of this century. For the last two decades, however, fertility rates have been fairly constant at about the replacement level. Chen and Morgan (1991) present evidence that the proportion of U.S. women who were childless at age 35 is very similar for women born in the early 1950s and women born during the late 19th century. What this evidence implies is that women are reducing their number of children but are not forgoing children all together. Instead, they are substituting quality for quantity.
The economic framework can also help explain other patterns of fertility and child care choice that we observe in the real world. For example, the historically high correlation between the presence of children and labor force participation has decreased over the last few decades. In 1950, labor force participation rates for married women with no children under 18 were almost three times higher than the rates for married women with children younger than 6 (Michael, 1985). In 1988, however, labor force participation rates for married women with no children under 18 were actually lower than the rates of married women with children younger than 6 (U.S. Bureau of the Census, 1990).[11] The growth in participation rates for women with young children has greatly exceeded the overall growth in participation rates of women. This observation is consistent with the hypothesis that before organized child care markets existed, there were few good substitutes for maternal care. A woman who didn't have any relatives nearby who were willing to provide child care would either stay at home to care for her children or choose not to have children at all. As better substitutes for mother's time became available in the market, women could choose to have children and work in the labor force.
Another example relates to the relationship between the work status of mothers and the quality of their children as measured by child developmental tests. The popular press has been quick to attribute the problems faced by children to the fact that more mothers are working. But academic studies have found that children are not, in general, worse off when the mother works and alternative forms of child care are used (Hayes et al., 1990). The household production approach can help explain why women who choose to work can make arrangements that are not detrimental to their children's well-being. As female wages have increased, women have been drawn into the labor force by better labor market opportunities. An increase in the mother's wage makes the family better off. If increases in income increase the demand for child quality, higher income families will demand higher quality children and will purchase higher quality care. This argument is less likely to hold for low-wage women, who are drawn into the labor market primarily because of low income rather than high wages. Unless government-subsidized child care programs are available, the only kind of child care that these women can afford is likely to be of lower quality.
Several aspects of Becker's approach have had profound impacts on the thinking of both economists and noneconomists, in particular, the recognition of the importance of time as an input to home-making and child-rearing activities and the concept of human capital. The household production approach models a sector of the economy that had received little attention from economists in the first half of the 20th century. Women's time was not valued (in economic terms), because it was primarily used in activities that were not included in national income accounts. By using the same kind of models for both firm production and household production, Becker's work legitimized the latter among mainstream economists and policymakers and made it clear that current measures of the Gross National Product (GNP) drastically underestimate economic output in the United States.[12] Some recent studies have calculated the underestimate to be as much as 20% to 40% (Blau & Ferber, 1992; Eisner, 1988).[13] The underestimate is even more severe if we believe that a correct measure of GNP should also include the value of investment in human capital, for example, time spent in school or parental time invested in their children. This recognition has led a number of leading economists to suggest that the definition of GNP be revised to account for household production and human capital investment, and both Canada and Australia have begun to revise their income accounts.
Human capital is a term that is now widely used by social scientists.[14] Investment in human beings enhances individual productivity and well-being and increases economic growth. In the 1960s and '70s, this idea was influential in shaping education and poverty policies, such as Head Start and Medicaid (Sawhill, 1988). In his work on the family, Becker broadened the concept of human capital to apply to more than just education and training. For example, individuals can invest in health or in a marriage or other kind of relationship. This concept also underlies the more recent debates about family values, with the recognition that the most important role of parents is investing in the next generation by instilling motivation (e.g., a work ethic) and other values.
The closely related idea of social capital developed by Coleman (1988) owes an intellectual debt to Becker. The theory of social capital expands on the idea of investing in relationships and includes relationships both inside and outside of the household. Social capital provides resources that increase productivity and well-being of families and individuals. For example, friends and relatives can provide help in an emergency or might provide a reference for a job. Neighborhood infrastructure such as parks, churches, libraries, clean streets, lack of violence, and so on can also be viewed as social capital. These investments can be made by individuals, families, or governments. Social capital theory will be discussed in greater depth in another article in this volume.
Becker argued that one of the advantages of the theory of household production over more standard economic theories of demand was that it allows us to separate the role of technology from that of tastes (Becker & Michael, 1973). This distinction is useful, because the implications for policy are quite different when differences in behavior are primarily driven by differences in tastes rather than differences in access to technology. For example, unwed teens may have babies because they want to (e.g., because of the need to feel wanted, grown up, etc.) or because of lack of knowledge or access to effective methods of fertility control. Similarly, some families may have less educated children because they have a preference for quantity over quality or because it is more expensive for these families to produce quality.[15]
One of the criticisms of household production models within economics disputes the possibility of separating tastes from technology. For technology to be independent of tastes, it must be true that the average cost of producing a given commodity is independent of other activities undertaken by the household and of the amount of the commodity produced. Pollak and Wachter (1975) pointed out that these two assumptions are not likely to hold for family production activities. For example, it is unlikely to take twice as much parental time for two children as for one, because parents can often be with both children at the same time. Therefore families who want to have a lot of children will have lower average costs of child rearing than families who want to have fewer children. The possibility of joint production will also make tastes and technology interdependent, because the costs of rearing children will be lower for families who enjoy the kind of activities that are easy to undertake with children. Although these issues make econometric estimation of household production functions more difficult, the household production framework and the critique outlined above are useful in clarifying our thinking and assessing the kinds of biases in estimates of the impact of various kinds of child care settings and/or parenting practices on the well-being of children.
One important contribution of the household production approach to interdisciplinary research is that it has allowed linkages to be developed between economists and noneconomists. In particular, the approach has given the standard measures of outcomes used by child development psychologists operational significance within economics. Economists use these outcome measures as dependent variables in models of child production, and child development psychologists provide insights into the nature of the production process. One example of the marriage between economists and develop mentalists is represented by the child data from the National Longitudinal Survey of Youth (NLSY).[16] The NLSY is a nationally representative data set of individuals who were 14 to 21 in 1979 and who have been surveyed annually since that time. The data are commonly used by economists and sociologists to study the school-to-work transition and the economic and demographic behavior of young adults. Since 1984 all the children of the female respondents have been given biannual assessments of cognitive, social, and emotional development using scales designed by child development psychologists. These data combine measures of child outcomes with information about the parents' characteristics and the amount and cost of parental inputs and market inputs, such as purchased child care. A data set that contains both economic and developmental data allows us to answer three interrelated questions:
1. What is the effect of various inputs on child outcomes (estimates of the available technology)?
2. How do prices of inputs (both time and market), family income, and technology affect the choice of child care (i.e., substitution in production)?
3. How does the cost of child care affect choices about the number of children and the amount of resources invested in children (substitution in consumption)?
ALTRUISM
The standard assumption about utility in traditional neoclassical economic models is that individuals are selfish. That assumption is useful when analyzing consumer demand in the market or other kinds of behavior in impersonal or competitive settings. For example, we would not expect an individual to say to a store clerk "Please, let me pay more for this product, because I know that you need the money." On the other hand, there are plenty of examples of nonselfish behavior, from charitable giving to helping an older person cross the street. The place where we most expect to find altruistic behavior is in the family, and Becker recognized that an analysis of the family would be incomplete without incorporating altruism.[17]
What are the advantages of explicitly modelling altruism? First, altruism between family members is more consistent with what we observe in the real world. The cost of children, especially in developed countries, greatly exceeds the economic benefits (Lindert, 1976). This could be true in an economic model only if the parents received positive utility from having children and from the children's well-being. Second, specifying these kinds of utility functions allows us to explicitly model behavior within the family. Third, altruism is what makes the family a unique social institution. Because of altruism, the family is better able than other institutions to perform certain functions such as instilling moral values and providing insurance against certain kinds of bad outcomes (Ben-Porath, 1980; Pollak, 1985). In this article, we classify models of altruism into three broad categories:
1. Parent-child altruism where the two parents have the same preferences toward their children
2. Husband-wife altruism
3. Parent-child altruism where the two parents have different preferences regarding their children.
A subset of this last category relates to children whose parents have divorced.
In Becker's (1974b) earliest work on altruism, he assumed that the family head (i.e., parent) is altruistic toward the children, but that the children are selfish.[18] For simplicity the model further assumed that there is only one parent or that the two parents pool their income and behave as one. This latter assumption has led some scholars to label Becker's model the common preference model (see, e.g., Thomas, 1990). The model is also referred to as the benevolent dictator model, because of the assumption that the parent controls family resources and cares equally about all children. The parent's goal is to maximize family income/resources and make transfers so that all children are equally well off. Altruistic parents can redistribute income to offset actions both inside and outside the family. For example, Becker's "Rotten Kid Theorem" says that the altruism of the parent will induce children to act altruistically, even if they are actually selfish. The rotten kid cannot gain at the expense of other family members, because the altruistic head will just reallocate income to offset the consequences of the rotten kid's behavior.
A different sort of offsetting behavior might exist when one child receives additional resources from outside the family from a government program such as Head Start. Because the altruistic head cares about all the family members, she or he would want everyone to share in the additional resources. This could be accomplished by reducing the family resources going to the child who is participating in the government program and increasing family resources going to other family members (including the altruist). Thus everyone in the family would be better off due to the government program, but the effect on the targeted child is smaller than if there were no offsetting reallocation of family resources. Becker suggested that this model could help explain why the measured effects of these kinds of child intervention programs are generally found to be small. An additional implication, however, is that the measured effects understate the full effect, which includes an increase in the well-being of all family members. A study by Seitz (1994), who found sibling effects of a parent-support program, is consistent with this argument.
The concept of parent-child altruism has also influenced macroeconomic debates about the intergenerational consequences of social security and the national debt. Barro (1974) pointed out that the children's generation is paying for their parent's social security through higher taxes. Altruistic parents will want to make transfers to their children to offset the higher tax burden caused by the existence of social security. The same argument holds for the intergenerational consequences of the national debt: altruistic parents will make transfers in the form of bequests or inter vivos gifts to offset the burden of debt left to their children's generation. A recent study projects that the baby boom generation will inherit trillion or ,000 per person (Avery & Rendall, 1993).
The model also has implications for what kinds of transfers will be made. If children differ in ability, the parent will make larger human capital investments in the more able, because there is a higher rate of return (i.e., earnings) on that kind of investment, and will make compensating financial transfers to the less able. That strategy is the most cost-effective way for the parent to increase all children's well-being. If parents follow this strategy, we would expect to see unequal monetary bequests across the children in the family. The evidence about the distribution of bequests is mixed. The majority of bequests are equal, but there is weak evidence that higher bequests go to offspring with lower income (Wilhelm, 1991). Why is the effect not stronger? One explanation has to do with incentive compatibility. Because it is sometimes difficult for parents to determine if the lower income is due to circumstances beyond their children's control or, instead, due to laziness, transferring more money to lower-income children reduces children's incentives to work hard and earn money on their own.
The issue of incentive compatibility and strategic behavior by nonal-truistic children is also relevant to the debate about the nature of parental altruism. Do the parents care about their children's well-being as assessed by the children themselves, or do parents have paternalistic preferences about particular outcomes?[19] If parents only care about children's well-being and not about particular outcomes, it is most cost-effective to make monetary transfers and let the children themselves decide how to spend the money. In reality, we often see transfers that are targeted toward one particular kind of consumption. Becker called these "merit goods." Parents will provide money to finance postsecondary education or a down payment on a house, but they will not give children money to bet at the horse track. Becker asserted this behavior is not necessarily due to paternalistic altruism but can be related to the possibility of strategic behavior on the part of the children. If parents care about the well-being of their children, it is not credible for the parents to give a certain dollar transfer at one point in time and promise to give no more, regardless of what might happen to their children at a later date. The children can recognize this fact and have less of an incentive to spend the money carefully, because they know that they can go back to their parents for more. If parents cannot distinguish between poor outcomes that are caused by careless spending and poor outcomes that are beyond the control of their children, they may prefer to make transfers in the form of merit goods to eliminate the possibility of this kind of strategic behavior.[20]
Altruism is also relevant to the current debate about welfare and what to do with children born to unmarried teenage mothers. Becker argues that because of parents' altruism, children are, in general, better off living with their parents than living in some other institution.[21] Altruistic parents would spend more on children out of each dollar of family income than would be spent by substitute caregivers. Thus, to attain the same level of child well-being, it is cheaper to pay poor parents to care for their own children than to pay unrelated individuals such as foster parents.[22] The fact that per-child payments to foster parents are twice as high as the level of benefits to parents from Aid to Families with Dependent Children (AFDC) is consistent with this idea (Avery, 1994).
The common preference altruism model is probably most useful in looking at families with young children. As children become older, they begin to have access to resources such as labor market earnings that are outside the parent's control. The question then becomes, How can parents maintain control of their children's behavior? In more traditional societies, where the extended family is more common, social norms reinforce the control of the family elder (Cheung, 1972). In agrarian societies, economics plays an important role as well. When land is the primary form of wealth, the family head can maintain control over that resource, even after the children are grown. In more modern societies bequests can play a similar role. One implication of the strategic role of bequests is that parents may invest less than optimally in the human capital of their children, because human capital investments come early in a child's life, and once invested, that leverage is gone.
Some economists have questioned whether bequests and other investments in children are altruistically motivated or whether, instead, an exchange model is more relevant to family behavior (Bernheim, Shleifer, & Summers, 1986; Cox, 1987). An example of this idea is the theory that parents have children to support them in old age. However, without social norms or reciprocal altruism, this kind of exchange is impossible to enforce. As above, rich parents could control their children's behavior through the threat of withholding bequests, but these parents do not need old-age support from their children. A second argument is that parents care less about monetary support than about having their children spend time with them. Rich parents could again enforce filial attention through bequests, but, without altruism or love, companionship from children would be no different than purchasing a companion in the market. These arguments imply that altruism must play an important role in the family.
Becker (1973, 1974a) and Becker, Landes, & Michael (1977) analyzed spousal relationships in a series of articles on marriage and divorce. These models relax two of the basic assumptions of the parent-child altruism models by allowing for both mutual altruism and spousal conflict when resources are limited and the preferences of each spouse differ. Marriage occurs not only because of sexual attraction and love (altruism), but also because marriage can provide additional benefits over being single through economies of scale (the per-person cost of living in a two-person household is cheaper than in a one-person household), coordination and sharing of child-rearing activities, and productivity gains from each spouse specializing in different tasks.[23 ] Becker described a marriage as a long-term implicit contract between the husband and wife. The contract specifies the roles for each spouse in producing the marriage output (e.g., market work, household tasks, etc.) and how the benefits from the marriage will be divided between the two. In other words, the implicit contract is a set of expectations about the interactions between the husband and wife.
The terms of the contract are decided in the marriage market. How the total marital resources are divided is a function of the desirability of each spouse and the alternatives (i.e., other prospects) available to each. Conflict over the distribution of resources is resolved in the marriage market, but there is no conflict over production: Once a division of the pie has been agreed on, the best strategy for both is to maximize total marital resources. An additional assumption in this model is transferrable utility. All this assumption means is that there is some tradeable commodity such that one spouse can compensate the other for undertaking activities that are undesirable. For example, the wife may agree to walk the dog for a week in return for the husband agreeing to invite her boss over for dinner.
Mutual altruism or love is important in marriage models because the state rarely intervenes to enforce implicit marriage contracts.[24]Altruism provides a reason to trust that the agreed-on division of resources will occur. Altruism can be thought of as a form of social capital that facilities the enforcement of the marriage contract. Altruism is not the only thing that ensures that the contract will be honored. The possibility of divorce can provide a minimum threshold for utility. Spouses will not want to stay in a marriage if they are worse off in that marriage than their alternatives at divorce. Whether the minimum level of utility within marriage is determined by alternatives in the original marriage market or by prospects at divorce depends on whether the law requires mutual consent for divorce to occur or allows for divorce on demand (Becker et al., 1977). Divorce on demand implies that for a marriage to continue, both spouses must be happier being married than being divorced. In related work, Peters (1986) provided evidence that mutual consent laws increase the bargaining power of women at divorce. She found that women who lived in mutual consent states had higher divorce settlements than women who lived in states with divorce on demand laws. The interpretation of this finding is that men who get divorced in mutual consent states must provide larger settlements to obtain their wives' consent, but divorce on demand laws allow them to leave without compensating their wives.
Family bargaining models developed by McElroy and Homey (1981) and Manser and Brown (1980) provide an alternative to Becker's framework. These articles use a cooperative bargaining framework first proposed by Nash (1953). In these models, each spouse has a threat point or default position that can be measured by prospects at divorce. These threat points provide limits on the outcomes that will be possible, because neither spouse will accept an outcome that is worse than their prospects at divorce. An implication of the models that distinguishes them from Becker's household dictator model is that observed outcomes should be a function of the relative bargaining power (i.e., the threat points) of each spouse. However, the implications of these bargaining models are actually very similar to Becker's models of marriage and divorce? Both assume that bargaining over marital resources allows the couple to maximize joint utility, and both recognize divorce as a threat point?
Bargaining models present the largest departure from Becker in analyzing parent/child relationships. If the two parents differ in their degree of altruism toward the children, then child outcomes will depend on who controls the family resources. It is often assumed that mothers are more altruistic toward their children than are fathers. If this is true, then children will be better off if child allowances or other child-related benefits are given to the mother (Lundberg & Pollak, 1993). Holding constant total family resources, Thomas (1990) found evidence from Brazilian data that children whose mothers had higher nonearned income were better off in terms of better nutrition, greater weight for height, and survival probabilities.
The issue of parental differences in altruism is also becoming important in developed countries as divorce, stepfamilies, out-of-wedlock births, and other nontraditional family structures become more common. It is often assumed that the low level of child support paid by noncustodial parents reflects a low degree of altruism toward their children. Alternatively, Weiss and Willis (1985) suggest that low payments are because the absent parent does not have any control over how the payments are spent. Peters, Argys, Maccoby, and Mnookin (1993) proposed a bargaining model in which some divorced parents can behave cooperatively, because each parent has control over something the other wants. The custodial parent controls access to the child and the noncustodial parent has control over monetary transfers. Low payments would then be expected from those absent parents who do not care about the threat of denied visitation.
Argys and Peters (1993) extended this bargaining model to explore how the implementation of state child-support guidelines--a formula for child support that is imposed unless the parents can agree to something else---can affect child-support awards and payments by altering the bargaining position of the custodial parent.
CONCLUSIONS
This article has described two areas of research in family economics that have been pioneered by Becker or have been extensions of his work or in response to his work. Household production models promise to address important questions such as how various child-rearing methods and circumstances affect child well-being and how costs, technological changes, and social changes affect the ways in which families rear children and, in turn, reflect the importance of children in our society. The topic of child rearing can be thought of much more broadly than just the choice of child care mode. The circumstances of child rearing also include family structure and other kinds of family background and relationships that are crucial inputs to child development. Until recently, household production models were more widely used in health economics, primarily because measures of health outcomes were available in national data sets. Adding child development measures to large nationally representative data sets used by economists has facilitated a dialogue and, perhaps, an eventual marriage between economists and child development psychologists.
Becker's work on altruism attempts to address the basic question of why parents invest in children. These models have let us begin to explore relationships and resource allocation within the family. The models have also provided insights into how parents' behavior toward their children is affected by government policies. The basic insight is that altruism produces links between family members. Thus anything that affects one family member can have effects on all family members, as gains or losses are spread throughout the family.
Becker's altruism model has come under some criticism because of the restrictiveness of the common preference assumption. This simple model, however, is useful in understanding how a little altruism can go a long way. We don't have to assume that everyone in the family is altruistic toward one another for altruism to have profound effects on family members' behaviors and resource allocation within the family. Relaxing the common preference assumption and allowing for the two parents to have different degrees of altruism toward children complicates the analysis of family interactions considerably. Introducing these complications, however, allows us to begin to address topics that are becoming increasingly important as divorce, stepfamilies, out-of-wedlock births, and other nontraditional family structures become more common.
* The work on this article was supported by NICHD Grant HD30944. The author thanks Keith Bryant, Jeff Evans, Cybele Raver, Michael Rendall, and William Schulze for comments.
NOTES
1. Much of the material that I discuss in this article is developed in Becker's (1991) Treatise on the Family and in his various earlier articles. Some of the earliest work on the economics of the family, inspired, in part, by Becker, is found in Schultz (1974). Willis (1987) provides a brief survey of the research in this area.
2. See Becker (1976) and Fuchs (1983) for further discussion of the economic approach.
3. Becker (1965) is the classic reference.
4. This proposition is similar to the political argument that promotes aggregate economic growth as a solution to poverty, that is "a rising tide carries all boats." In the case of the family, however, if family members are altruistic toward one another, there may be less controversy over the assumption that an increase in income is shared among all family members. In a later section, the article describes various models that relax this common preference/single decision maker assumption and focus on distributional issues.
5. For simplicity of exposition, the usual model assumes that the mother is the primary parental child care provider. The model could easily be extended to account for fathers who play a large role in child rearing. This extension would be particularly relevant because of recent evidence that shows an increase in the proportion of families who report the father as the primary child care provider (O'Connell, 1993).
6. Because women who stay at home also perform other household tasks, a more realistic comparison is to use the market cost of all the services that the mother performs: child care, house cleaning, laundry, shopping, bill paying, household management, and so on.
7. Social scientists often control for sociodemographic characteristics such as race/ethnicity, age, and urban/rural residence as a way to capture differences in tastes. Becker (1991) argued that some of these covariates may, in part, reflect differences in prices, opportunity costs, or technology. For example, historically higher fertility rates in rural areas may have reflected higher benefits of children due to their use as farm labor. Discrimination against Blacks in the labor market could lead to lower investments in the education of children and higher fertility rates (see the discussion about quantity/quality trade-offs later in this article). Economists interpret the error term in regression analysis as reflecting both measurement error and actual differences in behavior among otherwise observationally identical individuals. The latter effect can be interpreted as unobserved differences in tastes.
8. When mortality rates are high, it is risky to invest a lot of resources into children, because there is a substantial probability that the investment will be lost through the death of the child.
9. That is an issue that child development experts could address by measuring the improvement in day-care quality from a decrease in the child-staff ratio versus a comparable (in terms of cost) increase in educational toys and other equipment.
10. The assumption of equal investment is less likely to hold when the children in the family have very different innate abilities. For example, tastes for equality may lead parents to invest more resources in a child with disabilities. Parents may also invest more in a particularly gifted child, because the return on the investment is greater.
11. Part of the explanation for the lower rates of labor force participation for women with no children under 18 relates to cohort effects: This group includes older women who had lower levels of labor force participation throughout their life cycle.
12. Criticism of the exclusion of household production in national income accounts goes back at least as far as Kuznets (1941).
13. Note that the underestimate may have become smaller over time as women have increased their labor force participation and market-bought goods are substituted for home production. An offsetting factor is increases in home productivity due to technological change in the household sector and increases in the opportunity cost of women's time as market wages have increased. See Bryant and Zick (1994) for a treatment of household production that focuses on the allocation of time and its valuation.
14. Another Nobel prize winner, T. W. Schultz, was also recognized for his contributions to the development of human capital theory.
15. In Becker's (1975) book, Human Capital, he argued that poor families might invest in less education because it is more expensive to finance that education when there are no financial assets to serve as collateral for loans. Government financial aid serves as a means to overcome these capital market constraints.
16. See the NLSY Child Handbook (Baker, Keck, & Quinlan, 1993) for a more detailed description of these data.
17. More broadly, Becker has analyzed a number of situations where individuals' utilities are interdependent. Altruism exists when your well-being or consumption enters positively into my utility function: I am happy when you are better off. The converse is envy, where your well-being or consumption enters negatively into my utility function. A bandwagon effect exists when I am better off if my behavior conforms to that of others. This effect is similar to the concept of social norms used by sociologists or reference points used by psychologists. A bandwagon effect is a macrophenomenon that is one way to link individuals and households with the larger community in which they reside.
18. The framework of this model is very general and does not explicitly label the altruist as the parent and the selfish family member as the child, but the assumptions that altruism is unidirectional and that the altruist has control over the family resources are most relevant in describing parent-child relationships.
19. For a more in-depth discussion of this issue see Becker, 1991; Bruce and Waldman, 1990; and Pollak, 1988.
20. For very young children, the issue of parental control over transfers is probably less related to strategic behavior and more related to the child's inability to plan for the future. The important topic of how children mature and become able to make trade-offs between present and future consumption has not yet been adequately addressed in economic models of the family.
21. Of course there are cases of parental abuse and neglect where the parents are not acting altruistically. In those cases it is likely that the children would be better off in an alternative living arrangement. In addition, some parents are more altruistic than others. In general, however, it is not unreasonable to assume that parents are more altruistic toward their children than an individual selected at random would be.
22. This argument assumes that child well-being depends primarily on monetary expenditures. If we also consider issues of attachment, the importance of which is discussed in the child development literature, the argument for children to remain with their parents becomes even stronger.
23. Becker's model of the sexual division of labor within the household has generated a lot of criticism from those who argue that the model perpetuates sex-role stereotypes. The importance of specialization has undoubtedly decreased over time as more women have entered the labor market and as good substitutes for home production have increasingly become available in the marketplace (e.g., frozen gourmet dinners, wash and dry laundry services, etc.).
24. The state may intervene in cases of abuse and, at divorce, will also enforce prenuptial agreements about prior ownership of assets.
25. Grossbard-Shechtman (in press) makes a similar point.
26. A recent model developed by Lundberg and Pollak (1993) defines the threat point as a noncooperative marital outcome.
REFERENCES
Argys, L. M., & Peters, H. E. (1993, April). The role of child support guidelines in the negotiation of child support awards. Paper presented at the annual meetings of the Population Association of America, Cincinnati, OH.
Avery, R. (1994). Support of America's most vulnerable children: A policy analysis. Unpublished manuscript.
Avery, R. B., & Rendall, M. S. (1993). Estimating the size and distribution of baby boomers' prospective inheritance. In 1993 Proceedings of the Social Statistics Section, American Statistical Association (pp. 11-19).
Baker, P. C., Keck, C. K., & Quinlan, S. V. (1993). NLSY child handbook: A guide to the 1986-1990 National Longitudinal Survey of youth child data. Columbus, OH: Center for Human Resource Research.
Barro, R. J. (1974). Are government bonds net wealth? Journal of Political Economy, 82(6), 1095-1117.
Becker, G. S. (1965). A theory of the allocation of time. Economic Journal, 75(299), 493-517.
Becker, G. S. (1973). A theory of marriage: Part I. Journal of Political Economy, 81(4), 813-846.
Becker, G. S. (1974a). A theory of marriage: Part II. Journal of Political Economy, 82(2, pt.2), S11-S26.
Becker, G. S. (1974b). A theory of social interactions. Journal of Political Economy, 82(6), 1063-1093.
Becker, G. S. (1975). Human capital (2nd ed.). Chicago: University of Chicago Press.
Becker, G. S. (1976). The economic approach to human behavior. Chicago: University of Chicago Press.
Becker, G. S. (1991). A treatise on the family (enlarged ed.). Cambridge, MA: Harvard University Press.
Becker, G. S., Landes, E. M., & Michael, R. T. (1977). An economic analysis of marital instability. Journal of Political Economy, 85(6), 1141-1187.
Becker, G. S., & Michael, R. T. (1973). On the new theory of consumer behavior. Swedish Journal of Economics, 75, 378-395.
Ben-Porath, Y. (1980). The F-connection: Families, friends, and firms and the organization of exchange. Population Development Review, 6(1), 1-30.
Bernheim, B. D., Shleifer, A., & Summers, L. (1986). The strategic bequest motive. Journal of Labor Economics, 4(3, pt. 2), S151-S182.
Blake, J. (1989). Family size and achievement. Berkeley: University of California Press.
Blau, D. M., & Robins, P. K. (1988). Child care costs and family labor supply. The Review of Economics and Statistics, 70(3), 374-381.
Blau, F. D., & Ferber, M. A. (1992). The economics of women, men, and work. Englewood Cliffs, NJ: Prentice Hall.
Bruce, N., & Waldman, M. (1990). The rotten-kid theorem meets the samaritan's dilemma. Quarterly Journal of Economics, 105, 155-166.
Bryant, W. K., & Zick, C. D. (1994). The economics of housespousery: An essay on household work. Journal of Economic and Family Issues, 15(2), 137-168.
Chen, R., & Morgan, S. P. (1991). Recent trends in timing of first birth in the United States. Demography, 28(4), 513-534.
Cheung, S.N.S. (1972). The enforcement of property fights in children and the marriage contract. Economic Journal, 82(326), 641-657.
Coleman, J. S. (1988). Social capital in the creation of human capital. American Journal of Sociology, 94, S94-S120.
Connelly, R. (1992). The effect of child care costs on married women's labor force participation. The Review of Economics and Statistics, 75, 83-90.
Cox, D. (1987). Motives for private income transfers. Journal of Political Economy, 95(3), 508-546.
Datcher-Loury, L. (1988). Effects of mother's home time on children's schooling. The Review of Economics and Statistics, 70(3), 367-373.
Eisner, R. (1988) Extended accounts for national income and product. Journal of Economics Literature, 26(4), 1611-1684.
Fuchs, V. R. (1983). How we live. Cambridge, MA: Harvard University Press.
Grossbard-Shechtman, S. (in press). Marriage market models. In K. Ierulli and M. Tommasi (Eds.), The new economics of human behavior (pp. 92-112). Cambridge: Cambridge University Press.
Hayes, C. D., Palmer, J. L., & Zaslow, M. J. (1990). Who cares for America's children?: Child care policy for the 1990s. Washington, DC: National Academy Press.
Kuznets, S. (1941). National income and its composition, 1919-1938. New York: National Bureau of Economic Research.
Lindert, P. H. (1976). Child costs and economic development. In R. A. Easterlin (Ed.), Population and economic change in developing countries (pp. 5-79). Chicago and London: University of Chicago Press.
Lundberg, S., & Pollak, R. A. (1993). Separate spheres bargaining and the marriage market. Journal of Political Economy, 101(6), 988-1010.
Manser, M., & Brown, M. (1980). Marriage and household decision-making: A bargaining analysis. International Economic Review, 21, 31-44.
McElroy, M. B., & Horney, M. J. (1981). Nash-bargained household decisions: Toward a generalization of the theory of demand. International Economic Review, 22(2), 333-349.
Michael, R. T. (1985). Consequences of the rise in female labor force participation rates: Questions and probes. Journal of Labor Economics, 3(1, part 2), S117-S146.
Nash, J. F. (1953). Two-person cooperative games. Econometrica, 21, 128-140.
O'Connell, M. (1993). Where's papa? Fathers' role in child care. Population Trends and Public Policy, No. 20. Population Reference Bureau, Inc.
Peters, H. E. (1986). Marriage and divorce: Informational constraints and private contracting. American Economic Review, 76(3), 426-454.
Peters, H. E., Argys, L. M., Maccoby, E. E., & Mnookin R. H. (1993). Enforcing divorce settlements: Evidence from child support compliance and award modifications. Demography, 30(4), 719-735.
Pollak, R. A. (1985). A transaction cost approach to families and households. Journal of Economic Literature, 23(2), 581-608.
Pollak, R. A. (1988). Tied transfers and paternalistic preferences. American Economic Review, 78(2), 240-244.
Pollak, R. A., & Wachter, M. L. (1975). The relevance of the household production function and its implications for the allocation of time. Journal of Political Economy, 83(2), 255-277.
Sawhill, I. V. (1988). Poverty in the U.S.: Why is it so persistent? Journal of Economic Literature, 26(3), 1073-1119.
Schultz, T W. (1974). Economics of the family: Marriage, children, and human capital. Chicago: University of Chicago Press.
Seitz, V. (1994). Parent-focused intervention: Diffusion effects on siblings. Child Development, 65, 677-683.
Thomas, D. (1990). Intra-household resource allocation: An inferential approach. Journal of Human Resources, 25(4), 635-664.
U.S. Bureau of the Census. (1990). Statistical abstract of the United States: 1990 (110th ed.). Washington, DC: Government Printing Office.
Weiss, Y., & Willis, R. (1985). Children as collective goods in divorce settlements. Journal of Labor Economics, 3(3), 268-292.
Wilhelm, M. (1991). Bequest behavior and the effect of heirs earnings: Testing the altruistic model of bequests. Unpublished manuscript, University Park: The Pennsylvania State University.
Willis, R. J. (1987). What have we learned from the economics of the family? American Economic Review, 77(2), 68-81.
World Development Report, 1984 (pp. 122-123). New York: Oxford University Press.
