THE ECONOMIC WAY OF LOOKING AT LIFE

THE ECONOMIC WAY OF LOOKING AT LIFE* Nobel Lecture, December 9, 1992 G A R Y S. B E C K E R Department of Economics, University of Chicago, Chicago, IL. 60637, USA 1. The Economic Approach My research uses the economic approach to analyze social issues that range beyond those usually considered by economists. This lecture w ill describe the approach, and illustrate it with examples drawn from past and current work. Unlike Marxian analy sis, the eco no mic ap p ro ach I refer to d o es no t assume that individuals are motivated solely by selfishness or gain. It is a method of analysis, not an assumption about particular motivations. Along with others, I have tried to pry economists away from narrow assumptions about self interest. Behavior is driven by a much richer set of values and p references. The analysis assumes that individuals maximize welfare as they conceive it, w hether they be selfish, altruistic, lo yal, sp iteful, o r maso chistic. Their behavior is forward-looking, and it is also consistent over time. In particular, they try as best they can to anticipate the uncertain consequences of their actions. Forward-looking behavior, however, may still be rooted in the past, for the past can exert a long shadow on attitudes and values. Actions are constrained by income, time, imperfect memory and calculating capacities, and other limited resources, and also by the available opportunities in the eco no my and elsew here. These o p p o rtunities are larg ely determined by the private and collective actions of other individuals and organizations. Different constraints are decisive for different situations, but the most fund amental co nstraint is limited time. Eco no mic and med ical p ro g ress have greatly increased length of life, but not the physical flow of time itself, w hich alw ays restricts everyone to tw enty-four hours per day. So w hile goods and services have expended enormously in rich countries, the total time available to consume has not. Thus, wants remain unsatisfied in rich countries as well as in poor ones. * This lecture is dedicated to the memory of George J. Stigler, who died almost exactly one year ago. Nobel Laureate, outstanding economist, very close friend and mentor, he would be as happy as I am had he lived to see me deliver the 1992 Nobel Lecture in Economics. For while the growing abundance of goods may reduce the value of additional goods, time becomes more valuable as goods become more abundant. Utility maximization is of no relevance in a Utopia where everyone’s needs are fully satisfied , but the co nstant flo w o f time makes such a Uto p ia impo ssible. These are so me o f the issues analyzed in Becker [1965], and Linder [1970]. The follow ing sections illustrate the economic approach w ith four very different subjects. To understand discrimination against minorities, it is necessary to w iden preferences to accommodate prejudice and hatred of particular groups. The economic analysis of crime incorporates into rational behavior illegal and other antisocial actions. The human capital approach considers how the productivity of people in market and non-market situatio ns is changed by investments in educatio n, skills, and kno w ledge. The economic approach to the family interprets marriage, divorce, fertility, and relatio ns amo ng family members thro ugh the lens o f utility-maximizing forw ard-looking behavior. 2. Discrimination Against Minorities Discrimination against outsiders has always existed, but with the exception of a few discussions of the employment of w omen (see Edgew orth [1922], and Faucett [1918]), eco no mists w ro te little o n this subject befo re the 1950s. I began to w o rry abo ut racial, religio us, and gender discriminatio n w hile a graduate student, and used the concept of discrimination coefficients to organize my approach to prejudice and hostility to members of particular groups. Instead of making the common assumptions that employers only consider the productivity of employees, that w orkers ignore the characteristics of those with whom they work, and that customers only care about the qualities of the goods and services provided, discrimination coefficients incorporate the influence of race, gender, and other personal characteristics on tastes and attitudes. Emplo yees may refuse to w o rk under a w o man o r a black even when they are well paid to do so, or a customer may prefer not to deal w ith a black car salesman. It is o nly thro ugh widening of the usual assump tio ns that it is p o ssible to begin to und erstand the o bstacles to advancement encountered by minorities. Presumably, the amount of observable discrimination against minorities in wages and employment depends not only on tastes for discrimination, but also on other variables, such as the degree of competition and civil rights leg islatio n. Ho w ev er, asid e fro m the imp o rtant theo ry o f co mp ensating differentials originated by Adam Smith, and a few major studies like Myrdal’s American Dilemma [1944], there was little else available in the 1950s to build o n to analy ze ho w p rejud ice and o ther v ariables interact. I sp ent several years working out a theory of how actual discrimination in earnings and employment is determined by tastes for discrimination, along with the degree of competition in labor and product markets, the distribution of d iscriminatio n co efficients amo ng members o f the majo rity gro up , the 40 Economic Sciences 1992 access of minorities to education and training, the outcome of median voter and other voting mechanisms that determine whether legislation favors or is hostile to minorities, and other considerations. Since there was so much to be done in this field, my advisors encouraged me to convert my doctoral dissertation (Becker [1955]) into a book (Becker [1957]). The actual discriminatio n in the market place against a mino rity gro up depends on the combined discrimination of employers, w orkers, consumers, scho o ls, and g o v ernments. The analy sis sho w s that so metimes the environment greatly softens, while at other times it magnifies, the impact of a given amount of prejudice. For example, the discrepancy in w ages between equally productive blacks and whites, or women and men, would be much smaller than the degree of prejudice against blacks and women when many co mpanies can efficiently specialize in emplo ying mainly blacks o r w omen. Ind eed , in a w o rld w ith co nstant returns to scale in p ro d uctio n, tw o segregated economies with the same distribution of skills would completely bypass discriminatio n and w o uld have equal w ages and equal returns to o ther reso urces, regardless o f the desire to discriminate against the segregated mino rities. Therefo re, discriminatio n by the majo rity in the marketplace is effective because minority members cannot provide various skills in sufficient quantities to co mp anies that w o uld sp ecialize in using these w orkers. When the majority is very large compared to the minority - in the United States whites are nine times as numerous and have much more human and p hy sical cap ital p er cap ita than blacks - market d iscriminatio n by the majority hardly lowers their incomes, but may greatly reduce the incomes of the minority. However, when minority members are a sizable fraction of the total, discrimination by the majority injures them as well. This pro po sitio n can be illustrated w ith an analysis o f discriminatio n in South A frica, w here blacks are four to five times as numerous as w hites. Discriminatio n against blacks has also significantly hurt w hites, altho ugh so me w hite gro ups have benefitted (see Becker [1971, pages 30 - 31], and Hutt [1964]). Its sizable cost to w hites suggests w hy apartheid and other blatant forms of Afrikaaner discrimination eventually broke down. A literature has developed on whether discrimination in the marketplace due to prejudice disappears in the long run. Whether employers who do not want to discriminate will eventually compete away all discriminating employers depends not only on the distribution of tastes for discrimination among potential employers, but critically also on the nature of firm production functions. Of g reater sig nificance emp irically is the lo ng run d iscriminatio n by employees and customers, w ho are far more important sources of market discrimination than employers. There is no reason to expect discrimination by these groups to be competed away in the long run unless it is possible to have eno ugh efficient segregated firms and effectively segregated markets for goods. Gary S. Becker 41 A no vel theo retical develo pment in recent years is the analysis o f the co nsequences o f stereo ty p ed reaso ning o r statistical d iscriminatio n (see Phelps [1972], and Arrow [1973]). This analysis suggests that the beliefs o f employers, teachers, and other influential groups that minority members are less productive can be self-fulfilling, for these beliefs may cause minorities to underinvest in education, training, and work skills, such as punctuality. The underinvestment does make them less productive (see a good recent analysis by Loury [1992]). Ev id ence fro m many co untries o n the earning s, unemp lo y ment, and occupations of blacks, women, religious groups, immigrants, and others has expanded enormously during the past twenty-five years. This evidence more fully documents the economic position of minorities and how that changes in different environments. However, the evidence has not dispelled some of the controversies over the source of lower incomes of minorities (see Cain’s [1986] good review of both the theoretical literature and empirical analysis.) 3. Crime and Punishment I b eg an to think ab o u t c rim e in the 1960s af ter d riv ing to C o lu m b ia University for an oral examination of a student in economic theory. I was late and had to decide quickly whether to put the car in a parking lot or risk getting a ticket for parking illegally on the street. I calculated the likelihood of getting a ticket, the size of the penalty, and the cost of putting the car in a lot. I decided it paid to take the risk and park on the street. (I did not get a ticket.) A s I w alked the few blocks to the examination room, it occurred to me that the city authorities had probably gone through a similar analysis. The frequency of their inspection of parked vehicles and the size of the penalty imp o sed o n v io lato rs sho uld d ep end o n their estimates o f the typ e o f calculations potential violators like me w ould make. Of course, the first question I put to the hapless student was to work out the optimal behavior of both the offenders and the police, something I had not yet done. In the 1950s and 1960s intellectual discussions of crime were dominated by the opinion that criminal behavior was caused by mental illness and social oppression, and that criminals w ere helpless “ victims.” A book by a w ellknown psychiatrist was entitled The Crime of Punishment (see Menninger [1966]). Such attitudes began to exert a major influence on social policy, as laws changed to expand criminals’ rights. These changes reduced the apprehension and conviction of criminals, and provided less protection to the law-abiding population. I w as no t sympathetic to the assumptio n that criminals had radically different motivations from everyone else. I explored instead the theoretical and emp irical imp licatio ns o f the assump tio n that criminal behav io r is ratio nal (see the early pio neering w o rk by Bentham [1931] and Beccaria [1986]), but again “ ratio nality” did no t necessarily imply narro w materialism. It reco gnized that many peo ple are co nstrained by mo ral and ethical Economic Sciences 1992 42 considerations, and did not commit crimes even when they were profitable and there was no danger of detection. However, police and jails would be unnecessary if such attitudes always prevailed. Rationality implied that some individuals become criminals because of the financial rew ards from crime compared to legal w ork, taking account of the likelihood of apprehension and conviction, and the severity of punishment. The amo unt o f crime is d etermined no t o nly by the ratio nality and p references o f w o uld -be criminals, but also by the eco no mic and so cial environment created by public policies, including expenditures on police, p unishments fo r d ifferent crimes, and o p p o rtunities fo r emp lo yment, schooling, and training programs. Clearly, the type of legal jobs available as w ell as law , o rder, and punishment are an integral part o f the eco no mic approach to crime. Total public spending on lighting crime can be reduced, w hile keeping the mathematically expected punishment unchanged, by offsetting a cut in expenditures on catching criminals with a sufficient increase in the punishment to those convicted. How ever, risk-preferring individuals are more deterred from crime by a higher probability of conviction than by severe punishments. Therefore, optimal behavior by the State w ould balance the reduced spending on police and courts from low ering the probability of conviction against the preference of risk-preferring criminals for a lesser certainty of punishment. The State should also consider the likelihood of punishing innocent persons. In the early stages of my work on crime, I was puzzled by why theft is socially harmful since it appears merely to redistribute resources, usually from w ealthier to poorer individuals. I resolved the puzzle (Becker [1968, fn. 3] by recognizing that criminals spend on weapons and on the value of the time in planning and carrying out their crimes, and that such spending is socially unproductive - it is what is now called “ rent-seeking” - because it does not create wealth, only forcibly redistributes it. The social cost of theft was approximated by the number of dollars stolen since rational criminals would be willing to spend up to that amount on their crimes. (I should have ad d ed the reso urces sp ent by p o tential v ictims p ro tecting themselv es against crime.) One reason why the economic approach to crime became so influential is that the same analytic apparatus can be used to study enforcement of all laws, including minimum wage legislation, clean air acts, insider trader and other violations of security laws, and income tax evasions. Since few laws are self-enforcing, they require expenditures on conviction and punishment to deter violators. The United States Sentencing Commission has explicitly used the economic analysis of crime to develop rules to be follow ed by judges in punishing violators of Federal statutes (United States Sentencing Commission [1988]). Studies of crime that use the economic approach have become common d uring the p ast quarter century . These includ e analy sis o f the o p timal Gary S. Becker 43 marg inal p unishments to d eter increases in the sev erity o f crimes - fo r example, to deter a kidnapper from killing his victim (the modern literature starts w ith Stig ler [1970]), and the relatio n betw een p riv ate and p ublic enforcement of laws (see Becker and Stigler [1974], and Landes and Posner [1975]). Fines are p referable to imp riso nment and o ther ty p es o f p unishment because they are more efficient. With a fine, the punishment to offenders is also revenue to the State. The early discussions of the relations betw een fines and other punishments have been clarified and considerably improved (see, e.g., Polinsky and Shavell (19841, and Posner [1986]). Emp irical assessments o f the effects o n crime rates o f p riso n terms, conviction rates, unemployment levels, income inequality, and other variables have become more numerous and more accurate (the pioneering work is by Ehrlich [1973], and the subsequent literature is extensive). The greatest co ntro v ersies surro und the questio n o f w hether cap ital p unishment deters murders, a controversy that is far from being resolved (see, e.g., Ehrlich [1975], and National Research Council [1978]). 4. Human Capital Until the 1950s eco no mists generally assumed that labo r po w er w as given and no t augmentable. The so phisticated analyses o f investments in education and other training by Adam Smith, Alfred Marshall, and Milton Friedman w ere no t integrated into d iscussio ns o f p ro d uctiv ity. Then T. W . Schultz and others began to pioneer the exploration of the implications of human cap ital inv estments fo r eco no mic g ro w th and related eco no mic questions. Human capital analysis starts with the assumption that individuals decide on their education, training, medical care, and other additions to know ledge and health by weighing the benefits and costs. Benefits include cultural and other non-monetary gains along with improvement in earnings and occupations, while costs usually depend mainly on the foregone value of the time spent on these investments. Human capital is so uncontroversial nowadays that it may be difficult to appreciate the ho stility in the 1950s and 1960s to w ard the appro ach that went with the term. The very concept of human capital w as alleged to be demeaning because it treated people as machines. To approach schooling as an investment rather than a cultural experience w as co nsidered unfeeling and extremely narrow. As a result, I hesitated a long time before deciding to call my book Human Capital, and hedged the risk by using a long subtitle. Only gradually did economists, let alo ne o thers, ac c ep t the c o nc ep t o f human capital as a valuable tool in the analysis of various economic and social issues. My work on human capital began with an effort to calculate both private and social rates of return to men, w omen, blacks, and other groups from investments in different levels of education. A fter a w hile it became clear that the analysis of human capital could help explain many regularities in 44 Economic Sciences 1992 labor markets and the economy at large. It seemed possible to develop a more general theory of human capital that includes firms as well as individuals, and that could consider its macro-economic implications. The empirical analysis tried to co rrect data o n the higher earnings o f mo re educated perso ns fo r the fact that they are abler: they have higher I.Q.s and score better on other aptitude tests. It also considered the effects on rates of return to education of mortality, income taxes, foregone earnings, and economic grow th. A bility corrections did not seem very important, but large changes in adult mortality and sizeable rates of economic growth did have big effects. The empirical study o f investments in human capital received a majo r boost from Mincer’ s classic w ork [see 1974]. He extended a simple regression analysis that related earnings to years of schooling (Becker and Chisw ick [1966]) to includ e a crud e but v ery useful measure o f o n-the-jo b training and experience - years after finishing school; he used numerous individual observations rather than grouped data, and he carefully analyzed the pro perties o f residuals fro m earnings-generating equatio ns. There are now numerous estimated rates of return to education and training for many countries (for a summary of some of this literature, see Psacharopoulos [1975]). The accumulating evidence on the economic benefits of schooling and training also promoted the importance of human capital in policy discussions. This new faith in human capital has reshaped the way governments approach the problem of stimulating growth and productivity, as was shown by the emphasis on human capital in the recent presidential election in the United States. One of the most influential theoretical concepts in human capital analysis is the distinctio n betw een general and specific training o r kno w ledge (see Becker [1962], and Oi [1962]). By d efinitio n, firm-sp ecific kno w led g e is useful only in the firms providing it, whereas general knowledge is useful also in o ther firm s. Teac hing so m eo ne to o p erate an IBM -c o m p atib le perso nal co mputer is general training, w hile learning the autho rity structure and the talents o f emp lo yees in a p articular co mp any is sp ecific knowledge. This distinction helps explain why workers with highly specific skills are less likely to quit their jobs and are the last to be laid off during business dow nturns. It also explains w hy most promotions are made from w ithin a firm rather than thro ug h hiring - w o rkers need time to learn about a firm’ s structure and “ culture” - and why better accounting metho d s w o uld includ e the sp ecific human cap ital o f emp lo yees amo ng the principle assets of most companies. Firm-specific investments produce rents that must be shared betw een employers and employees, a sharing process that is vulnerable to “ opportunistic” behavior because each side may try to extract most of the rent after investments are in place. Rents and opportunism due to specific investments play a crucial role in the modern economic theory of organizations (see Williamson [1985]), and in many discussions of principal-agent prob- Gary S. Becker 45 lems (see, fo r examp le, Gro ssman and Hart [1983]). The imp licatio ns o f specific capital for sharing and turnover have also been used in analyzing marriage “ markets” to explain divorce rates and bargaining within a marriage (see Becker, Landes and Michael [1977], and McElro y and Ho rney [1981]), and in analyzing political “ markets” to explain the low turnover of po liticians (see Cain, Ferejo hn and Firo ina [1987]). The theory of human capital investment relates inequality in earnings to differences in talents, family background, and bequests and other assets (see Becker and Tomes [1986]). Many empirical studies of inequality also rely on human capital concepts, especially differences in schooling and training (see Mincer [1974]). The sizeable gro w th in earnings inequality in the United States during the 1980s that has excited so much political discussion is largely explained by higher returns to the more educated and better trained (see, e.g., Murphy and Welch [1992]). Human capital theory gives a provocative interpretation of the so-called “ gender gap” in earnings. Traditionally, women have been far more likely than men to work part-time and intermittently partly because they usually withdrew from the labor force for a while after having children. As a result, they had fewer incentives to invest in education and training that improved earnings and job skills. During the past twenty years all this changed. The decline in family size, the growth in divorce rates, the rapid expansion of the service sector where mo st w o men are emp lo yed , the co ntinuing eco no mic d ev elo p ment that raised the earnings o f w o men alo ng w ith men, and civil rights legislatio n encouraged greater labor force participation by w omen, and hence greater investment in market-oriented skills. In practically all rich countries, these forces significantly improved both the occupations and relative earnings of w omen. The United States’ experience is especially well-documented. The gender gap in earnings amo ng full-time men and w o men remained at abo ut 35 percent from the mid-fifties to the mid-seventies. Then w omen began the steady economic advance which is still continuing; it narrowed the gap to und er 25 p ercent. W o men are flo cking to business, law , and med ical schools, and are working at skilled jobs that they formerly shunned, or were excluded from. Schultz and others (see, e.g., Schultz [1963], and Denison [1962]) early on emphasized that investments in human capital were a major contributor to economic growth. But after a while the relation of human capital to growth was neglected, as economists became discouraged about whether the available growth theory gave many insights into the progress of different countries. The revival of more formal models of endogenous growth has brought human cap ital o nce again to the fo refro nt o f the d iscussio ns (see e.g., Ro mer [1986], Lucas [1988], Barro and Sala-i-Martin [1992], and Becker, Murphy and Tamura [1990]). 46 Economic Sciences 1992 5. Formation, Dissolution, and Structure of Families The rational choice analysis of family behavior builds on maximizing behavior, investments in human capital, the allocation of time, and discrimination against w omen and other groups. The rest of the lecture focuses on this analysis since it is still quite controversial. Writing A Treatise on the Family is the most difficult sustained intellectual effort I have undertaken. The family is arguably the most fundamental and oldest of institutions - some authors trace its origin to more than 50,000 years ago. The Treatise tries to analyze not only modern Western families, but also those in other cultures and the changes in family structure during the past several centuries. Trying to cover this broad subject required a degree of mental commitment over more than six years, during many nighttime as well as daytime hours, that left me intellectually and emotionally exhausted. In his autobiography, Bertrand Russell says that writing the Principia Mathematica used up so much of his mental powers that he was never again fit for really hard intellectual w ork. It took about tw o years after finishing the Treatise to regain my intellectual zest. The analysis o f fertility has a lo ng and ho no rable histo ry in eco no mics, but until recent y ears marriag e and d iv o rce, and the relatio ns betw een husbands, wives, parents, and children had been largely neglected by economists (although see the important study by Mincer [1962]). The point of departure of my work on the family is the assumption that when men and w o men d ecid e to marry o r hav e child ren o r d iv o rce, they attemp t to maximize their utility by comparing benefits and costs. So they marry when they expect to be better off than if they remained single, and they divorce if that is expected to increase their welfare. People w ho are not intellectuals are often surprised w hen told that this approach is controversial since it seems obvious to them that individuals try to raise their welfare by marriage and divorce. The rational choice approach to marriage and other behavior is in fact often consistent with the instinctive economics “ of the common man” (Farrell and Mandel [1992]). Still, intuitive assumptions about behavior is only the starting point o f systematic analysis, fo r alo ne they do no t yield many interesting implicatio ns. The ratio nal cho ice ap p ro ach embed s them in a framew o rk that combines maximizing behavior with analysis of marriage and divorce markets, specialization and the division of labor, old-age support, investments in children, and legislation that affects families. The implications of the full mo d el are o ften no t so o bv io us, and so metimes run sharp ly co unter to received opinion. For example, contrary to a common belief about divorce among the rich, the economic analysis of family decisions shows that wealthier couples are less likely to divorce than poorer couples. A ccording to this theory, richer co uples tend to gain a lo t fro m remaining married, w hereas many po o rer couples do not. A poor woman may well doubt whether it is worth staying married to someone chronically unemployed. Empirical studies for many Gary S. Becker 47 countries do indicate that the marriages of richer couples are much more stable (see Becker, Landes and Michael [1977]). Efficient bargaining between husbands and wives implies that the trend in Europe and the United States toward no-fault divorce during the past two decades would not raise divorce rate and, therefore, that contrary to many claims, it could not be responsible for the rapid rise in these rates. However, the theory does indicate that no-fault divorce hurts w omen w ith children whose marriages are broken up by their husbands. Households headed by unmarried women with children now comprise about one-fifth of all households with children in the United States and other advanced countries. Economic models of behavior have been used to study fertility ever since Malthus’ s classic essay; the great Sw edish economist, Knut Wicksell, w as attracted to economics by his belief in the Malthusian predictions of overpopulation. But Malthus’ s conclusion that fertility w ould rise and fall as inco mes increased and decreased w as co ntradicted by the large decline in birth rates after some countries became industrialized during the latter part of the nineteenth century and the early part of this century. The failure of Malthus’ s simple model of fertility persuaded economists that family-size decisions lay beyond economic calculus. The neo-classical grow th model reflects this belief, for in most versions it takes population grow th as exogenous and given (see, for example, Cass [1965], or A rrow and Kurz [1970]). Ho w ev er, the tro uble w ith the Malthusian ap p ro ach is no t its use o f eco no mics p er se, but an eco no mics inap p ro p riate fo r mo d ern life. It neg lects that the time sp ent o n child care beco mes mo re exp ensiv e as countries become more productive. The higher value of time raises the cost of children, and thereby reduces the demand for large families. It also fails to consider that the greater importance of education and training in industrialized economies encourages parents to invest more in the skills of their children, which also raises the cost of large families. The growing value of time and the increased emphasis o n scho o ling and o ther human capital explain the decline in fertility as countries develop, and many features of birth rates in modern economies. Why in almo st all so cieties have married w o men specialized in bearing and rearing children and in certain agricultural activities, whereas married men have do ne mo st o f the fighting and market w o rk? The explanatio n, presumably, is a combination of biological differences betw een men and w o men - especially differences in their innate capacities to bear and rear children - and discriminatio n against w o men in market activities, partly through cultural conditioning. Large and highly emotional differences of opinion exist over the relative importance of biology and discrimination in generating the traditional division of labor in marriages (see, for example, Boserup [1970]). The economic analysis of this division of labor does not determine the relative importance of biology and discrimination, but it shows how sensitive the division is to small d ifferences in either. Since the return fro m 48 Economic Sciences 1992 investing in a skill is greater when more time is spent utilizing the skill, a married couple could gain a lot from a sharp division of labor because the husband could specialize in some types of human capital and the w ife in others. Given such a large gain from specialization within a marriage, only a little discrimination against women or small biological differences in childrearing skills w ould cause the division of labor betw een household and market tasks to be systematically related to gender. The sensitivity to small d ifferences exp lains w hy the emp irical ev id ence canno t read ily cho o se between biological and “ cultural” interpretations. This theory also explains w hy many w o men entered the labo r fo rce as families became smaller, divorce more common, and earning opportunities for w omen improved. Relations among family members differ radically from those among emp lo y ees o f firms and members o f o ther o rg anizatio ns. The interactio ns betw een husbands, w ives, parents, and child ren are mo re likely to be motivated by love, obligation, guilt, and a sense of duty than by self-interest narrow ly interpreted. It was demonstrated about twenty years ago that altruism within families eno rmo usly alters ho w they resp o nd to sho cks and p ublic p o licies that redistribute resources among members. Becker [1974] show ed that exogenous redistributions of resources from an altruist to her beneficiaries (or vice versa) may not affect the welfare of anyone because the altruist would try to reduce her gifts by the amount redistributed. Barro [1974] derived this result in an intergenerational context, which cast doubt on the common assumptio n that go vernment deficits and related fiscal po licies have real effects on the economy. The “ Ro tten-Rid Theo rem” - the name is very popular even when critics disagree w ith the result - carries the analysis of altruism further, for it shows how the behavior of selfish individuals is affected by altruism. Under so m e c o nd itio ns, ev en selfish b enefic iaries - o f c o u rse, m o st p arents believe that the best example of this is selfish children with altruistic parents - are induced to act as if they are altruistic to w ard their benefacto rs because that raises their ow n selfish w elfare. They act this w ay because otherw ise gifts from their benefactors w ould be reduced enough to make them worse off (see Becker [1974], and the elaboration and qualifications to the analysis in Lindbeck and Weibull [1987], Bergstrom [1989], and Becker [1991, p p . 9-13]). The Bible, Plato’s Republic, and other early writings discussed the treatment of young children by their parents, and of elderly parents by adult children. Bo th the elderly and children need care - in o ne case because o f declining health and energy, and in the other because of biological growth and dependency. A pow erful implication of the economic analysis of relations within families is that these two issues are closely related. Parents who leave sizable bequests do not need old-age support because instead they help o ut their children. I mentio ned earlier o ne w ell-kno w n imp licatio n o f this: und er certain co nd itio ns, bud get d eficits and so cial security payments to the elderly have no real effects because parents simply Gary S. Becker 49 o ffset the big g er taxes in the future o n their child ren thro ug h larg er bequests. It is much less appreciated that altruistic parents who leave bequests also tend to invest mo re in their children’ s skills, habits, and values. Fo r they gain from financing all investments in the education and skills of children that yield a higher rate o f return than the return o n savings. They can ind irectly sav e fo r o ld ag e by inv esting in child ren, and then red ucing bequests when elderly. Both parents and children would be better off when parents make all investments in children that yield a higher return than that on savings, and then adjust bequests to the efficient level of investment (see section 1 of the Appendix for a formal demonstration). Even in rich countries many parents do not plan on leaving bequests. These parents w ant o ld-age suppo rt, and they “ underinvest” in their child ren’ s ed ucatio n and o ther care. They und erinv est because they canno t compensate themselves for greater spending on children by reducing bequests since they do not plan on leaving any. Both the children and parents would be better off if the parents agreed to invest more in the children in return for a commitment by the children to care for them w hen they need help. But how can such a commitment be enforced? Economists and lawyers usually recommend a written contract to insure co mmitment, but it is absurd to co ntemp late that a so ciety w ill enforce contracts betw een adults and ten-year-olds or teenagers. Part of my current research considers an indirect w ay to generate commitments w hen pro mises and w ritten agreements are no t binding. I w ill d escribe briefly so me o f this new w o rk because it carries the eco no mic ap p ro ach to the family unto uncharted g ro und related to the ratio nal formation of preferences w ithin families. Parental attitud es and behav io r hav e an eno rmo us influence o n their children. Parents who are alcoholic or are addicted to crack create a bizarre atmo sp here fo r imp ressio nable y o ung sters, w hereas p arents w ith stable values w ho transmit kno w ledge and inspire their children favo rably influence both what their children are capable of and what they want to do. The economic approach can contribute insights into the formation of preferences thro ug h child ho o d exp eriences w itho ut necessarily ad o p ting the Freudian emphasis o n the primacy o f w hat happened during the first few months of life. A g ain, I am try ing to mo d el a co mmo n sense id ea; namely , that the attitudes and values of adults are enormously influenced by their childhood experiences. A n Indian doctor living in the United States may love curry because he acquired a strong taste for it w hile grow ing up in India, or a woman may forever fear men because she was sexually abused as a child. Through its assumption of forward-looking behavior, the economic point of view implies that parents try to anticipate the effect of what happens to children on their attitudes and behavior w hen adults. These effects help determine the kind o f care parents pro vide. Fo r example, parents w o rried about old-age support may try to instill in their children feelings of guilt, 50 Economic Sciences 1992 obligation, duty, and filial love that indirectly, but still very effectively, can “ commit” children to helping them out. Economists have too narrow a perspective on commitments. “ Manipulating” the experiences of others to influence their preferences may appear to be inefficient and fraught with uncertainty, but it can be the most effective w ay to obtain commitment. Economic theory needs to incorporate guilt, affection, and related attitudes into preferences in order to have a deeper understanding o f w hen co mmitments are “ credible” (see section 2 of the appendix for a formal discussion). Parents who do not leave bequests may be willing to make their children feel guiltier precisely because they gain more utility from greater old-age consumption than they lose from an equal reduction in children’s consumption. This type of behavior may be considerably more common than suggested by the number of families that leave bequests, for parents w ith young children o ften do no t kno w w hether they w ill be financially secure w hen they are old. They may try to protect themselves against ill health, unemplo yment, and o ther hazards o f o ld age by instilling in their children a willingness to help out if that becomes necessary. This analysis of the link between childhood experiences and adult preferences is closely related to work on rational habit formation and addictions (see Becker and Murphy [1988]). The formation of preferences is rational in the sense that parental spending on children partly depends on the anticipated effects of childhood experiences on adult attitudes and behavior. I do not have time to consider the behavior of children - such as crying and acting “ cute” - that tries in turn to influence the attitudes of parents. Many economists, including myself, have excessively relied on altruism to tie together the interests of family members. Recognition of the connection betw een childhood experiences and future behavior reduces the need to rely on altruism in families. But it does not return the analysis to a narrow focus on self-interest, for it partially replaces altruism by feelings of guilt, obligation, anger, and other attitudes usually neglected by models of rational behavior. If p arents anticip ate that child ren w ill help o ut in o ld ag e - p erhap s because of guilt or related motivations - even parents w ho are not very lo ving to w ard their children w o uld invest mo re in the children’ s human capital, and save less to provide for their old age. (For a proof, see section 3 of the Appendix.) Equation (12) of the A ppendix show s that parents alw ays prefer small increases in their own consumption to equal increases in their children’s if the only way they can get greater consumption is by making children feel guiltier. This means that altruistic parents w ho take steps to make their children feel guiltier alw ays underinvest in the children’ s human capital. This shows directly why creating guilt has costs and is not fully efficient. Altruistic family heads who do not plan to leave bequests try to create a “ warm” atmosphere in their families, so that members are willing to come to the assistance of those experiencing financial and other difficulties. This Gary S. Becker 51 conclusion is relevant to discussions of so-called “ family values,” a subject that receiv ed attentio n d uring the recent p resid ential camp aig n in the United States. Parents help determine the values of children - including their feelings of obligation, duty, and love - but what parents try to do can be greatly affected by public policies and changes in economic and social conditions. Consider, for example, a program that transfers resources to the elderly, p erhap s esp ecially to p o o rer families w ho d o no t leav e bequests, that red uces the eld erly ’ s d ep end ence o n child ren. A cco rd ing to the earlier analysis, parents who do not need support when they become old do not try as hard to make children mo re lo yal, guiltier, o r o therw ise feel as w elldisposed toward their parents. This means that programs like social security that significantly help the elderly would encourage family members to drift apart emotionally, not by accident but as maximizing responses to those policies. Other changes in the mo dern w o rld w hich have altered family values include increased geographical mobility, the greater wealth that comes with eco no mic g ro w th, better cap ital and insurance markets, hig her d iv o rce rates, smaller families, and publicly-funded health care. These developments have generally made peo ple better o ff, but they also w eakened the personal relations within families between husbands and wives, parents and children, and amo ng mo re distant relatives, partly by reducing the incentives to invest in creating closer relations. 6. Concluding Comments An important step in extending the traditional analysis of individual rational choice is to incorporate into the theory a much richer class of attitudes, preferences, and calculations. This step is prominent in all the examples that I co nsider. The analysis o f discriminatio n includes in preferences a d islike o f - p reju d ic e ag ainst - members of particular groups, such as blacks or women. In deciding whether to engage in illegal activities, potential criminals are assumed to act as if they consider both the gains and the risks - including the likelihood they will be caught and severity of punishments. In human capital theory, people rationally evaluate the benefits and co sts o f activ ities, such as ed ucatio n, training, exp end itures o n health, migration, and formation of habits that radically alter the way they are. The economic approach to the family assumes that even intimate decisions like marriage, divorce, and family size are reached through weighing the advantages and disadvantages of alternative actions. The weights are determined by preferences that critically depend on the altruism and feelings of duty and obligation toward family members. Since the economic, or rational choice, approach to behavior builds on a theory of individual decisions, criticisms of this theory usually concentrate o n p articular assump tio ns abo ut ho w these d ecisio ns are mad e. A mo ng o ther things, critics deny that individuals act co nsistently o ver time, and question whether behavior is forward-looking, particularly in situations that 52 Economic Sciences 1992 differ significantly from those usually considered by economists - such as those involving criminal, addictive, family, or political behavior. This is not the place to go into a detailed response to the criticisms, so I simply assert that no approach of comparable generality has yet been developed that offers serious competition to rational choice theory. While the economic approach to behavior builds on a theory of individual choice, it is not mainly concerned w ith individuals. It uses theory at the micro level as a powerful tool to derive implications at the group or macro level. Rational individual choice is combined with assumptions about technologies and other determinants of opportunities, equilibrium in market and nonmarket concerning the implications at those studying situations, and laws, norms, and traditions to obtain results behavior of groups. It is mainly because the theory derives the macro level that it is of interest to policymakers and differences among countries and cultures. N o ne o f the theo ries co nsid ered in this lecture aims fo r the g reatest generality; instead, each tries to derive concrete implications about behavior that can be tested w ith survey and other data. Disputes over w hether punishments deter crime, whether the lower earnings of women compared to men are mainly due to discrimination or lesser human capital, or whether no-fault divorce law s increase divorce rates all raise questions about the empirical relevance of predictions derived from a theory based on individual rationality. A close relation between theory and empirical testing helps prevent both the theo retical analysis and the empirical research fro m beco ming sterile. Empirically oriented theories encourage the development of new sources and types of data, the way human capital theory stimulated the use of survey data, especially panels. A t the same time, puzzling empirical results force changes in theory, as models of altruism and family preferences have been enriched to cope with the finding that parents in Western countries tend to bequeath equal amounts to different children. I have been impressed by how many economists want to work on social issues rather than issues forming the traditional core of economics. At the same time, specialists from fields that do consider social questions are often attracted to the economic way of modelling behavior because of the analytical pow er provided by the assumption of individual rationality. Thriving schools of rational choice theorists and empirical researchers are active in sociology, law, political science, history, anthropology, and psychology. The rational choice model provides the most promising basis presently available for a unified approach to the analysis of the social world by scholars from the social sciences. A PPEN D IX 1. To show this formally, suppose that each person lives for three periods: youth (y), middle age (m), and old age (o), and has one child at the beginning of period m. A child’s youth overlaps his parents’ middle age, and a child’s Gary S. Becker 53 middle age o verlaps his parents’ o ld age. The utility p arents get fro m altruism is assumed to be separable from the utilities produced by their own consumption. A simple utility function of parents (V,) incorporating these assumptions is Each p erso n w o rks and earns inco me o nly d uring mid d le ag e. It is possible to save then to provide consumption for old age (Z o p ) by accumulating assets w ith a y ield o f R k . Parents influence child ren’ s earning s by investing in their human capital. The marginal yield o n investments in human capital (Rh) is defined as (2) w here Ec is the earnings of children at middle age. This yield is assumed to decline as more is invested in children: dRt,/ dh I 0, where h is the amount invested. Parents must also d ecid e w hether to leav e bequests, d eno ted by k,. If p arents can co nsume at d ifferent ag es, leav e bequests, o r inv est in the child’s human capital, their budget constraint is where A is the present value of resources. One first o rder co nditio n to maximize parental utility determines their optimal consumption at middle and old age and the last determines investments in the human capital of children Equatio n (6) assumes that the first o rd er co nd itio n fo r inv estment in human capital is a strict equality; that some human capital is always invested in children. This can be justified w ith an Inada-type condition that small investments in human capital yield very high rates of return. In rich economies like Sweden or the United States, investments in basic knowledge and 54 Economic Sciences 1992 nutrition of children presumably do yield a very good return. A s long as p arents are no t c o m p letely self ish - as lo ng as a > 0 - then su c h a co nd itio n d o es alw ay s imp ly p o sitiv e inv estment in human cap ital. Fo r completely selfish parents, equation (6) would be an inequality. Equation (4) determines the accumulation of assets to finance old-age co nsump tio n. W hether p arents leav e bequests o r w ant o ld -age sup p o rt from their children is determined by the inequality in (5). If this is a strict inequality, parents want support and would not leave bequests. That inequality can be written in a more revealing way. If children also maximize their utility, then the envelope theorem implies that (7) Equation (7) has the intuitive interpretation that parents do not give bequests when the utility they get from their children consuming a dollar more at middle age is less than the utility they get from a dollar more of their own consumption at old age. Obviously, such an inequality holds for completely selfish parents since the left-hand side of equations (5) and (7) are zero when a is zero . The w eaker the altruism (the smaller a) the mo re parents w ant from children. Combining equations (5) and (6) gives (8) Equatio n (8) imp lies that the marg inal rate o f return o n human cap ital equals the return o n assets w hen parents give bequests, and it is greater than the asset return when parents do not give bequests. Parents can help children either by investing in their human capital or by leaving them assets. Since they w ant to maximize the advantage to children, given the cost to themselv es - p arents are no t sad istic - they help in the mo st efficient fo rm. Consequently, if strict inequality holds in equation (8), they w ould not give bequests, for the best way to help children when the marginal return on human capital exceeds that on assets is to invest only in human capital. Parents leave bequests only when they get the same marginal return on both (some of these results have been derived in Becker and Tomes [1986]). 2. To analyze in a simple way the influence of parents over the formation of child ren’ s p references, sup p o se p arents can take actio ns x and y w hen children are young that affect children’s preferences when adults. I use the assump tio n o f sep arability to w rite the utility functio n o f mid d le-aged children as I assume that H’ > 0 and G x > 0, which means that an increase in y raises the utility of children, but an increase in x lowers their utility. For concreteness, interpret H as “ happiness,” and G as the “ guilt” children feel toward Gary S. Becker 55 their parents, so that greater x makes children feel guiltier. The question is: why would non-sadistic parents want to make their children feel guilty? The v ariable g is the key to und erstand ing w hy . This measures the contribution of children to the old-age support of parents; let us assume that c hild ren f eel less g u ilty w hen they c o ntrib u te m o re (Gs < 0). If G, > 0, then greater x bo th raises children’ s guilt and stimulates mo re giving by them. The budget constraint of parents becomes: The first-order condition for the optimal y is The first order condition for x is more interesting, for even altruistic parents may want to make their children feel guilty if that sufficiently raises old-age support. This first order condition can be written as where dG/ dx incorporates the induced change in g. The second term in the middle expression is negative to altruistic parents because greater x does raise child ren’ s guilt, w hich lo w ers the utility o f these p arents (a > 0). However, guilt also induces children to increase old-age support, as given by dg/ dx. The magnitude of this response determines whether it is worthwhile for parents to make children feel guiltier. Increased o ld -age sup p o rt fro m child ren has tw o p artially o ffsetting effects on the welfare of altruistic parents. On the one hand, it raises their o ld -ag e co nsump tio n and utility , as g iv en by u&,. On the other hand, it lowers children’s consumption, and hence the utility of altruistic parents, as given by -auk,. This means that altruistic parents who leave bequests never try to make children feel guiltier, for u&, = auk, for these parents. Since dG/ dx > 0, they must be worse 05 when their children feel guiltier. 3. Combine the first order conditions in Equations (5) and (6) to get (13) Both sides of this equation exceed unity when parents do not give bequests. Since greater old-age support from children low ers the left-hand side by lo w ering the numerato r and raising the deno minato r, the right-hand side must also fall to be in a utility maximizing equilibrium. But since Rk is given by market co nditio ns, the right-hand side can fall o nly if R h, falls, w hich 56 Economic Sciences 1992 implies greater investment in children when parents expect greater old-age support from children. Even completely selfish parents (a=0) might invest in children if that w ould sufficiently increase the expected old-age support from guilty children. BIBLIO G RA PH Y Arrow, Kenneth J., (1973) “The Theory of Discrimination,” in Orley Ashenfelter and Albert Rees, eds., Discrimination in Labor Markets (Princeton NJ: Princeton University Press), pp. 3-33. Arrow, Kenneth J. and Kurt, Mordecai, (1970) Public Investment, the Rate of Return, and Optimal Fiscal Policy (Baltimore: Published for Resources for the Future by Johns Hopkins University Press). Barro, Robert J., (1974) “Are Government Bonds Net Wealth?” Journal of Political Economy 82: 1095 - 1117. Barro, Robert J. and Sala-i-Martin, X., (1992) “Convergence,” Journal of Political Economy 100: 223-51. Beccaria, Cesarc, marchese di, (1986) On Crimes and Punishment, 1st ed., translation of Dei Delith e delle Pene (Indianapolis: Hackett Publishing Co.). Becker, Gary S., (1955) “Discrimination in the Market Place,” Ph.D. dissertation, University of Chicago. Becker, Gary S., (1957) The Economics of Discrimination, 1st ed. (Chicago: University of Chicago Press). Becker, Gary S., (1962) “Investment in Human Capital: A Theoretical Analysis,” Journal of Political Economy LXX: 9 - 49. Becker, Gary S., (1964) Human Capital, 1st ed. (New York: Columbia University Press for the National Bureau of Economic Research). Becker, Gary S., (1965) “A Theory of the Allocation of Time,” Economic ]ournal 75 (299):493-517. Becker, Gary S., (1968) “Crime and Punishment: An Economic Approach,“ Journal of Political Economy 76: 169 - 2 17. Becker, Gary S., (1971) The Economics of Discrimination, 2nd ed. (Chicago: University of Chicago Press). Becker, Gary S., (1974) “A Theory of Social Interactions,” Journal of Political Economy 82: 1063 -93. Becker, Gary S., (1991) A Treatise on the Family, expanded ed. (Cambridge MA: Harvard University Press). Becker, Gary S. and Chiswick, Barry, (1966) “Education and the Distribution of Earnings,” American Economic Review LVI: 358 - 69. Becker, Gary S., Landes, Elisabeth M., and Michael, Robert T., (1977) “An Economic Analysis of Marital Instability," ]ournal of Political Economy 85: 1141 -87. Becker, Gary S. and Murphy, Kevin M., (1988) “The Family and the State,” Journal of Law and Economics 3 1: 1 - 18. Becker, Gary S., Murphy, Kevin M., and Tamura, Robert, (1990) “Human Capital, Fertility, and Economic Growth,” Journal of Political Economy 98: S12 - 37. Becker, Gary S. and Stigler, George J., (1974) “Law Enforcement, Malfeasance, and Compensation of Enforcers,” Journal of Legal Studies 3: 1 - 18. Becker, Gary S. and Tomes, Nigel, (1986) “Human Capital and the Rise and Fall of Families,” Journal of Labor Economics 4: S 1 -S39. Bentham, Jeremy, (1931) Theory of Legislation (New York: Harcourt, Brace). Bergstrom, Theodore, (1989) “A Fresh Look at the Rotten-Kid Theorem - And Other Household Mysteries," Journal of Political Economy 97: 1 138 - 59. Boserup, Ester, (1970) Women’s Role in Economic Development (London: Allen and Unwin). Gary S. Becker 57 Cain, Bruce E., Ferqjohn, John, and Fiorina, Morris, (1987) The Personal Vote: Constituency Service and Electoral Independence (Cambridge MA: Harvard University Press). Cain, Glen G., (1986) “The Economic Analysis of Labor Market Discrimination: A Survey,” in O. Ashenfelter and R. Layard, eds., Handbook of Labor Economics, Handbooks in Labor Economics Series, no. 5 (New York: Elsevier Science): 693785. Cass, David, (1965) “Optimal Growth in an Aggregative Model of Capital Accumulation,” Review of Economic Studies 32:233-240. Denison, Edward F., (1962) Sources of Economic Growth in the United States (Washington, D.C.: Committee for Economic Development). Edgeworth, F. Y., (1922) “Equal Pay to Men and Women for Equal Work” Economic Journal XXXII: 431-57. Ehrlich, Isaac, (1973) “Participation in Illegitimate Activities: A Theoretical and Empirical Investigation,” Journal of Political Economy 81: 521-65. Ehrlich, Isaac, (1975) “The Deterrent Effect of Capital Punishment: A Question of Life and Death,” American Economic Review 85: 397-417. Farrell, C. and Mandel, M., “Uncommon Sense,” Business Week, October 26, 1992, pp. 36-37. Faucett, M., (1918) “Equal Pay for Equal Work,” Economic Journal XXVIII: l-6. Grossman, Sanford J. and Hart, Oliver D., (1983) “An Analysis of the PrincipalAgent Problem,” Econometrica 51: 7 - 45. Hutt, William H., (1964) The Economics of the Colour Bar: A Study of the Economic Origins and Consequences of Racial Segregation in South Africa (London: Published for the Institute of Economic Affairs by A. Deutsch). Landes, William M. and Posner, Richard A., (1975) “The Private Enforcement of Law," Journal of Legal Studies 4:1-46. Lindbeck, Assar and Weibull, Jorgen W., (1988) “Strategic Interaction with Altruism: The Economics of Fait Accompli,” Journal of Political Economy 96: 1165-82. University of Stockholm. Linder, Staffan Burenstam, (1970) The Hurried Leisure Class (New York: Columbia University Press). Loury, Glenn C., (I992) “Incentive Effects of Affirmative Action,” Annals of the American Academy of Political and Social Science 523: 19 - 29. Lucas, Robert E., Jr., (1988) “On the Mechanics of Economic Development,” Journal of Monetary Economics 22: 3-42. McElroy, Marjorie B. and Horney, Mary Jean, (198 1) “Nash-Bargained Household Decisions: Toward a Generalization of the Theory of Demand,” International Economic Review 22: 333-49. Menninger, Karl, (1966) The Crime of Punishment (New York: The Viking Press). Mincer, Jacob, (1962) "Labor Force Participation of Married Women,” in Aspects of Labor Economics, a conference of the Universities- National Bureau Committee for Economic Research (Princeton NJ: Princeton University Press for the National Bureau of Economic Research). Mincer, Jacob, (I974) Schooling, Experience, and Earnings (New York: Columbia University Press for the National Bureau of Economic Research). Murphy, Kevin M. and Welch, Finis, (I992) “The Structure of Wages,” Quarterly Journal of Economics 107: 285 - 326. Myrdal, Gunnar, (1944) An American Dilemma: The Negro Problem and Modern Democracy, 2 vo1s. (New York: Randon House). National Research Council (U.S.), Panel of Research on Deterrent and Incapacitative Effects, (1978) Deterrence and Incapacitation: Estimating the Effects of Criminal Sanctions on Crime Rates, Alfred Blumstein, Jacqueline Cohen, and Daniel Nagin, eds. (Washington, D.C.: National Academy of Sciences). 58 Economic Sciences 1992 Oi, Walter Y., (1962) “Labor as a Quasi-Fixed Factor,” Journal of Political Economy 70: 538-55. Phelps, Edmund S., (1972) “The Statistical Theory of Racism and Sexism,” American Economic Review 62: 659 - 61. Polinsky, A. Mitchell and Shavell, Steven, (1984) “The Optimal Use of Fines and Imprisonment,” Journal of Public Economics 24: 89 -99. Posner, Richard A., (1986) Economic Analysis of Law, 3rd ed. (Boston: Little, Brown). Psacharopoulos, George, (1975) Earnings and Education in OECD Countries (Paris: Organization for Economic Co-operation; Washington, D.C.: OECD Publications Center). Romer, Paul M., (1986) “Increasing Returns and Long Run Growth,” Journal of Political Economy 94: 1002-37. Schultz, Theodore W., (1963) The Economic Value of Education (New York: Columbia University Press). Stigler, George J., (1970) “The Optimum Enforcement of Laws,” Journal of Political Economy 78: 526 - 36. United States Sentencing Commission (1988) “Discussion Draft of Sentencing Guidelines and Policy Statement for Organizations: Proposed Chapter 8 for the Guidelines Manual.” Williamson, Oliver E., (1985) The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting (New York: Free Press; London: Collier Macmillan). A CKN O W LED GEM EN TS 1 have had valuable comments from James Coleman, Richard Posner, Sherwin Rosen, Raaj Sah, Jose Scheinkman, Richard Stern, and Stephen Stigler.