deeply into its potential implications for the central problems of fertilityand female labor-force participation. The timing and spacing of children,the opportunities for part-time work and accumulation of lifetime labormarket experience, and choices as to the amount of education to beinvested in early in the life cycle all revolve heavily on the terms underwhich women can participate in the labor market and thus share in thetransformation of the household's time resources into market commodities.The human capital literature (Ben-Porath 1967; Mincer 1970, 1974b) is,of course, replete with dynamic analyses of investment in human capitalover time and the life-cycle effects of these investments on earnings. Butlittle of this work has entered the more general framework of the "newhome economics," particularly as this theory bears on decisions concerningthe numbers of children and their timing and spacing within a marriageand the relation of these decisions to the accumulation of other forms ofassets. The work of Heckman (1971), Ghez and Becker (1972), and un-published work of Frank Stafford does, however, represent a notablebeginning of the extension of this part of theory into more dynamicallyrelevant realms.8 Yet we need to understand far more than we presentlydo about why the labor market functions so differently for men than forwomen, the role of institutional constraints, discrimination, and therelation of these to women's choices of occupation and timing of labor-forceparticipation. Once again, the simultaneities of the system severelylimit our ability to break out a single segment for proper analysis.

The fourth and final element in the new home economics is the resourceconstraints facing the household in its production and optimizationdecisions. Traditionally these constraints are divided into time (husband'sand wife's, although often the husband is assumed to devote full time tothe market) and "other" nonwage income. While it is universally recog-nized that some elements of household production and consumption—sleep and food, for example—are in fact inputs into the production-of-time resources, little attention has thus far been paid to the quality of thetime resources and of other family resources—both genetic and material—passed from one generation to the next. Arleen Leibowitz's study ofTerman's 1921 sample of California school children, reported in this

1'

MARC NERLOVE POPULATION AND ECONOMIC GROWTH 535

n is often the beneficiaryal that takes place in the1971, pp. 3 1—32) suggestsder of a child, may be an)robabilities, presumablyand nutrition that take

reasons.of separability of

the basically two-goodusually considered, forrent outputs in a multi-ions of variable propor-Ut of one good could bether good and without

production unit, in thisg resources fully. In aible for several of thealthough substitutable

the latter case, holdingI of one of the outputsher, although clearly ane can easily see how

could be in a contextr example, health andnt. In a dynamic con-en greater significancerly underinvestment inproductivity of later

973, pp.is a set of assumptionsincipally time, can beies to be used in the

I suppose one couldal technology of house-element of the theory.

y, since most of what isId members can entersomewhat secondarily,rchased. It is here that

economics cuts most

factors" and vice versa, and

desirable in future work toFor an empirical example,

Much recent work, however, is dynamic only in the relatively trivial sense of in-volving maximization over a number of periods of time without uncertainty concerningthe values of future values of exogenous constraints to future decisions. This is the sensein which a dynamic programming problem can be turned into an ordinary programmingproblem of much larger dimensions; apart from the computational difficulties thusintroduced, the chief defect of this approach is that it fails to lay bare the sequentialnature of the decision-making process in the way, for example, in which the recursivesolution to the general dynamic programming problem of Bellman does. In the absenceof uncertainty, however, nothing essential is lost by the straightforward multiperiodextension of the basically static framework. When values of future constraints are Un-certain, then it does become essential to understand and incorporate the sequentialcharacter of the decision process (see Nerlove 1 972a). So far this has not been done in theliterature of the new home economics.

536 MARC NERLOVE

volume, bears importantly on the manner in which much human capitalis passed from one generation to another, especially to a child of preschoolage. The sample is very unrepresentative, but it is instructive, for theinvestments in her child of a mother's time and the quality of thoseinvestments, as measured by her education, are found to affect appreciablylater measures of the child's ability and future earning capacity.

The resource constraints facing the household, once it is formed, are,of course, a product of the household formation itself and thus connectedwith my earlier remarks on the inseparability of this complex of issues,from those of family choice and decision making. But, more importantit is in this area that the complex issue of intergenerational transfersfigures most prominently. We live, after all,

in a world where individuals are born naked, destitute, helpless,ignorant, and untrained, and must spend a third of their livesin acquiring the prerequisites of a free contractual existence.The fundamental fact about society as a going concern is thatit is made up of individuals who are born and die and give placeto others; and the fundamental fact about modern civilizationis that it is dependent upon the utilization of three great accu-mulating funds of inheritance from the past, material goods andappliances, knowledge and skill, and morale. Besides the torchof life itself, the material wealth of the world, a technologicalsystem of vast and increasing intricacy and the habituationswhich fit men for social life must in some manner be carriedforward to new individuals born devoid of all these things asolder individuals pass out. The existing order, with the insti-tutions of the private family and private property (in self as wellas goods), inheritance and bequest and parental responsibility,affords one way for securing more or less tolerable results ingrappling with this problem. [Knight 1921, pp. 374—75]

So the apparently simple theoretical construct of a time budget plusother income constraint to the household conceals beneath its sereneand mathematically differentiable exterior the central problem of thecontinuity of society itself.9

1.POPULATION AND ECON

By themselves the founew home economics—not physical commodit(2) a household produenvironment providinginto market commodities.are incapable of yieldinmain problems of houseonly a frameworkspecial additional assumust be added to thepropositions. Moreover,ations is intimately reladata to which the newrange all the way fromof time, to cross-sectionregions at a point in tibased on individual mttive information.

If to the key simpliindependent productivassumptions that yountime in comparisonchildren are less intensithe comparative advamarket-purchasable copartner, a number of iare tolerably well supSetting aside for the moto measure the costimplication of thefamily will cause a suchildren and toward tcommodities. Indeed,completely specializerepresent pure incomefamilies with working.income and negativethat education servesmeasure some of these cfrom market wages, siof male and female edSchultz 1970; and pavolume).

The interpretation

Closely connected to the matter of what is inherited from the past and transferredfrom generation to generation is the problem of how individual tastes and preferencesare formed and how they may change over time. I have not mentioned this issue lest Ibe excommunicated from the economics profession! It is virtually part of the definitionof what an economist is that he takes tastes as given, and I sometimes suspect that manyof us require all tastes to be identical and assume that all differences among individualsarise from differences in the resource constraints those individuals face. Indeed, one ofthe consequences of the conjectures and speculations presented in Section 3 is that thetastes and preferences of at least different generations, if not of individuals within eachgeneration, could remain constant while the number of children per family declinedover time due to changes in the value of human time induced endogenously.

MARC NERLOVE POPULATION AND ECONOMIC GROWTH 537

By themselves the four main elements of the theoretical structure of thenew home economics—(l) a utility function with arguments which arenot physical commodities but home-produced bundles of attributes;(2) a household production technology; (3) an external labor-marketenvironment providing the means for transforming household resourcesinto market commodities; and (4) a set of household resource constraints—are incapable of yielding a series of well-defined implications about themain problems of household behavior with which we are concerned. It isonly a framework within which to think about these problems. Manyspecial additional assumptions, some of which have been mentioned,must be added to the framework to arrive at empirically refutablepropositions. Moreover, the nature of the required additional specific-ations is intimately related to the peculiarities of the particular bodies ofdata to which the new home economics has been applied. These datarange all the way from aggregate time-series data covering long periodsof time, to cross-sectional census data for both large and small geographicregions at a point in time and over time, to household and family databased on individual interviews, with and without collection of retrospec-tive information.

If to the key simplification involved in the assumption of separableindependent productive processes within the household, one adds theassumptions that young children are highly intensive of the mother'stime in comparison with other activities within the home and olderchildren are less intensive, and that for institutional or biological reasonsthe comparative advantage of the male partner in the acquisition ofmarket-purchasable commodities significantly exceeds that of the femalepartner, a number of interesting implications of the theory emerge whichare tolerably well supported by the empirical evidence so far analyzed.Setting aside for the moment the inadequacy of the observed market wageto measure the cost of a nonworking woman's time, the immediateimplication of the theory is that a rise in cost of mother's time for thefamily will cause a substitution away from time-intensive goods such aschildren and toward those requiring more inputs of market-purchasablecommodities. Indeed, if we further assume momentarily that males arecompletely specialized in market activities, changes in their wagesrepresent pure income effects for the family, and we do then observe forfamilies with working mothers positive association of family size withincome and negative association with female wage rates. To the extentthat education serves as a proxy for the relative costs of time which maymeasure some of these costs less imperfectly, or at least in a fashion differentfrom market wages, similar differences are observed between the effectsof male and female educational attainments (Mincer 1963; Nerlove andSchultz 1970; and papers by Willis, DeTray, and Ben-Porath in thisvolume).

The interpretation of many of the results is somewhat complicated

ich much human capitalEly to a child of preschoolit is instructive, for thend the quality of thoseund to affect appreciablyarning capacity.I, once it is formed, are,Ltself and thus connectedf this complex of issues,g. But, more importanttergenerational transfers

aked, destitute, helpless,nd a third of their livesontractual existence.a going concern is that

n and die and give place'out modern civilizationion of three great accu-past, material goods andriorale. Besides the torch

world, a technologicaly and the habituations

manner be carriedLd of all these things asg order, with the insti-property (in self as wellparental responsibility,

less tolerable results in921, pp. 374—75]

of a time budget pluseals beneath its serenecentral problem of the

the past and transferredidual tastes and preferences

mentioned this issue lest IFtually part of the definition

suspect that manyamong individuals

face. Indeed, one ofin Section 3 is that the

of within eachFdldren per family declinedN endogenously.

538 MARC NERLOVE

by the effects of marriage on the association of the educational attain-ments of husbands and wives, and by the fact that wives and mothers forthe most part specialize entirely in home activities during only varyingfractions of their lifetimes. Moreover, the effects of female education seemunexpectedly and puzzlingly nonlinear (see Ben-Porath's paper herein).To illustrate the nature of the difficulty involved, consider the followingsimple analysis of the effects of assortive mating. We know that men ofhigher-than-average levels of educational attainment tend to marry womenof higher-than-average levels of educational attainment. Suppose, how-ever, that tastes for children differ among the population; women with ahigh taste for children and a low taste for market-related activities will tendin general to seek and to receive less formal schooling than women with theopposite preference. Presumably a man's taste for children in a societywhere nearly all men specialize in market-related activities will not bereflected very greatly in his formal schooling. Some effect of the amountof his formal schooling may occur through early marriage, but he willcertainly seek a mate with like preferences. Thus, men with a given levelof schooling with a high taste for children will tend to marry women withless schooling than the average associated with the level these men haveachieved. If, as is common, a husband's educational attainment isassociated primarily with permanent income effects within the household,and his wife's is associated primarily with the opportunity cost of time,one can see that the negative effect of the opportunity cost of time asmeasured this way on fertility will tend to be exaggerated, holding maleeducational attainment constant, since the difference between the twoacross couples partly reflects differences in tastes which are unobservableand not included in the statistical analysis (see also "The EmergingEconomic Picture" in Part I of T. W. Schultz's paper in this book for arelated comment).

Turning the analysis around, we can ask what implications the newhome economics has for female labor-force participation. The new homeeconomics predicts what is perhaps the obvious: the composition of awoman's family is strongly associated with her labor-force participation.Typically, the number of a family's children under the age of 18 and theage of the youngest child are both strong predictors of a woman's labor-force participation. These facts about family composition have beeninterpreted by Cain (1966), for example, as measures of the opportunityvalue of a mother's time in the home. Gronau (herein) attempted a de-tailed and sophisticated analysis of just this proposition, using a subsampleof the 1/1000 sample from 1960 United States Census; moreover, heexamined the interaction of educational attainment with family com-position. Others (Smith 1972a; Leibowitz 1972) have documented thathighly educated married women participate to a greater extent in thelabor force and work more hours when they do work than married womenwith less schooling. Married women as a group also tend to withdraw

•1 * 1 .POPULATION AND ECON

from the labor force wthe new home econofemale time—intensivehome; but the rates atments withdraw is nothighly educated womethan do women with IU-shaped relation betIsraeli women with yoeffects of education atand market productioand child quality. Butthe presently existing F(1968), Nerlove and Scsized, educational invesbe viewed, at least pasurely one of the most iand only partly negate

Finally, if we regard,home economics as thethe "economic and socithe marked decline frolow birth and death rChildren"), the newlimited, to offer (0'mortality affects the cnumber of children sthe whole parents acthe longer a child surrelative increases in thages to successivelycording to the newchildren. This need noldeclines in mortalityof numbers of childre!technology of thewell as on the relative.factors, of course,culture, and placemust surely constituteapplication of the nevhowever, requires thatmodated in a model 9which we now turn.

MARC NERLOVE POPULATION AND ECONOMIC GROWTH 539from the labor force when they have children; this is an implication ofthe new home economics, on the assumption that children are morefemale time—intensive than other commodities produced within thehome; but the rates at which women with different educational attain-ments withdraw is not the same. During the child-rearing years, morehighly educated women reallocate more hours to household productionthan do women with less education. Ben-Porath's finding (herein) of aU-shaped relation between education and labor-force participation forIsraeli women with young children strongly suggests differences in theeffects of education at different levels on the relative efficiencies of homeand market production and also interactions between female educationand child quality. But all of this has been insufficiently explored withinthe presently existing framework of the new home economics. As Sweet(1968), Nerlove and Schultz (1970), and Hall (in this book), have empha-sized, educational investments, labor-force participation, and fertility mustbe viewed, at least partly, as simultaneously determined choices. This issurely one of the most important implications of the new home economicsand only partly negated by the latter's currently static character.

Finally, if we regard, as I think we must, the grand problem of the newhome economics as the explanation of the demographic transition, that is,the "economic and social processes and family behavior that accounts forthe marked decline from very high birth and death rates to modern verylow birth and death rates" (T. W. Schultz, in Part I, "The Value ofChildren"), the new home economics does have some insights, albeitlimited, to offer (O'Hara 1972). Clearly, a high probability of childmortality affects the costs of achieving a given family size, that is, thenumber of children surviving to a given age. If it is assumed that onthe whole parents achieve a greater (discounted) sum of satisfactionsthe longer a child survives and if declines in mortality result in greaterrelative increases in the conditional probabilities of survival from earlierages to successively older ages, declines in mortality should tend, ac-cording to the new home economics, to generate a greater demand forchildren. This need not lead to an increase in births, however, since suchdeclines in mortality lower the cost of child quality relative to the costof numbers of children. Of course, the net effect must depend on thetechnology of the production of child numbers and child quality aswell as on the relative importance of these in the utility function. Thesefactors, of course, may vary substantially from time to time, culture toculture, and place to place. The elucidation of such effects, however,must surely constitute one of the central challenges to the empiricalapplication of the new home economics. To explore such effects fully,however, requires that the household decision-making process be accom-modated in a model of economic growth and development, a subject towhich we now turn.

the educational attain-wives and mothers for

ities during only varyingof female education seemn-Porath's paper herein).

consider the followingg. We know that men ofient tend to marry womentainment. Suppose, how-opulation; women with arelated activities will tendling than women with thefor children in a society

activities will not beeffect of the amount

marriage, but he willmen with a given level

nd to marry women withhe level these men havecational attainment ists within the household,pportunity cost of time,rtunity cost of time asggerated, holding malerence between the twowhich are unobservablee also "The Emergingpaper in this book for a

implications the newipation. The new home

the composition of aparticipation.

the age of 18 and theof a woman's labor-

have beenures of the opportunityerein) attempted a de-tion, using a subsampleCensus; moreover, heent with family corn-have documented that

greater extent in thethan married women

Iso tend to withdraw

540 MARC NERLOVE POPULATION AND ECONI

3. Household Decision Making and Economic Growth:Speculations and Conjectures

"Would you tell me, please, which way I ought to go fromhere?"

"That depends a good deal on where you want to get to,"said the Cat. [Alice's Adventures in Wonderland]

In his classic paper on "Diminishing Returns from Investment,"Knight (1944) pointed out that "if new investment can be freely directedto all uses, i.e., embodied in all types of productive agents indifferently, itwill not be subject to diminishing returns" (p. 33). Moreover, he stressed,as did Marshall before him, the concept of capital in human beings, andthat "in the production of laborers the matter of 'quality' is far moreimportant than that of quantity in the crude sense of numbers" (p. 35).Although investment "freely directed" might not be subject to diminishingreturns, certainly, under static circumstances, continued investment inany one particular direction ought eventually to result in a declining rateof return. Yet, as T. W. Schultz has emphasized, the rate of returnappears to have diminished little, if at all, in response to a high leveland even accelerated pace of investment in human capital, and, indeed,it may have actually risen (T. W. Schultz 1971, p. 173).

We do not, of course, have any really accurate measure of the extent ofinvestment in human capital as compared with investment in nonhumancapital and in the stock of knowledge through investment in researchand development. Yet, there are a number of clues which suggest that thecapital stock invested in human beings, even on a per capita basis, hasbeen a steadily growing portion of the total capital stock. T. W. Schultz(1961, p. 73) suggests a rise of the value of the stock of educational humancapital embodied in the stock of labor of persons age 14 and older from18 percent in 1900 of the total educational and physical capital to 30percent in 1957. These estimates do not include on-the-job training orinvestments in better health and nutrition. In terms of gross capitalformation, Kuznets (1966, p. 243) calculates, on the basis of Schultz'searlier work, a rise in the share of investments in formal education alonefrom "about 9% in 1900 to over 38% in the 1950s." Moreover, forWestern countries as a whole, Kuznets (1971) calculates the share oflabor has risen from 55 to 75 percent of national income over the sameperiod. These facts, meager as they may be, suggest two significant ques-tions which are germane to the issue with which this paper began, namely,how can the new home economics be integrated into a general theory ofeconomic growth and development in a manner which has some hope,at least, of bearing on the grand question of the demographic transition?These two questions are the following.

First, what accountsin human capital, evena high and acceleratinto other forms? In othdisequilibrium among

Second, quite apartreturn (even if they arecapital investment perthe household and whgrowth of populationconjectures on the athe role of the increasiwhich is reciprocal, tbeings.

Razin (1969, 1972) sof human to total capgrowth path of an ecoearlier, T. W. Schultzof rates of return to mvfor skills and knowlebecause of additionaltechnique (tèchnologicand adjustment of thestock and seek out thmaintain the growthdisequilibrium createthat the rates of returfall only slowly, in revestment because ofthat is, changes on thd

One of the most mlhuman beings ascapita, as pointedof human time perincreasing value of di(1970) in his penetraland Schultz fail to nolincreasing value of huicapital as well as tocal change, and theplace. If one assumes,quality and quantity)produced within the

MARC NERLOVE

mic Growth:

way I ought to go from

re you want to get to,"erland]

ms from Investment,"nt can be freely directede agents indifferently, it

). Moreover, he stressed,1 in human beings, andof 'quality' is far morese of numbers" (p. 35).e subject to diminishing

ontinued investment inesult in a declining rateed, the rate of return

esponse to a high levelan capital, and, indeed,

• 173).easure of the extent of

vestment in nonhumaninvestment in researchwhich suggest that the

a per capita basis, hasal stock. T. W. Schultz

of educational humanage 14 and older fromphysical capital to 30on-the-job training orterms of gross capitalthe basis of Schultz's

formal education alone• 1950s." Moreover, for

alculates the share ofincome over the sametwo significant ques-

s paper began, namely,a general theory of

has some hope,transition?

POPULATION AND ECONOMIC GROWTH 541

First, what accounts for the failure of the rate of return to investmentsin human capital, even counting educational investments, to fall, despitea high and accelerating rate of investment in this form of capital relativeto other forms? In other words, why does there appear to be a persistentdisequilibrium among these rates of return?

Second, quite apart from the possibility of disequilibrium rates ofreturn (even if they are in equilibrium), what effect will increasing humancapital investment per capita have on the allocation of resources withinthe household and what, if any, repercussions will it have for the rate ofgrowth of population and labor force? A number of speculations andconjectures on the answers to these questions follow; they emphasizethe role of the increasing value of human time over time and its relation,which is reciprocal, to the increasing level of investment in humanbeings.

Razin (1969, 1972) showed how, under certain circumstances, the ratioof human to total capital per capita would increase along the optimalgrowth path of an economy experiencing technical progress. As indicatedearlier, T. W. Schultz has emphasized this aspect of the persistant failureof rates of return to investments in human capital to decline. The demandfor skills and knowledge embodied in human capital does not declinebecause of additional investments in the stock of useful knowledge andtechnique (technological change) which require the continual adaptationand adjustment of the human agent to utilize efficiently this augmentedstock and seek out the new sources of investment opportunities whichmaintain the growth process. But even in the absence of a persistentdisequilibrium created by the demand for human capital, it is possiblethat the rates of return to such investments would fail to fall over time, orfall only slowly, in relation to the rates of return to other forms of in-vestment because of endogenously changing relative cost of investment,that is, changes on the supply side.

One of the most important consequences of the growing "quality" ofhuman beings as reflected in the increased stock of human capital percapita, as pointed out by T. W. Schultz herein, is the increasing valueof human time per unit of such time. Many of the consequences of theincreasing value of time over time are amusingly explored by Linder(1970) in his penetrating study of Tue Harried Leisure Class. Yet Linderand Schultz fail to note the important link which may exist between theincreasing value of human time, due presumably to investment in humancapital as well as to investment in other forms of capital and in technologi-cal change, and the terms on which investment in human capital takesplace. If one assumes, as I think plausible, that children (as regards bothquality and quantity) are time-intensive as compared with other goodsproduced within the home, it follows that unless the increasing investment

.

542 MARC NERLOVE

of human capital increases the marginal productivity of a unit of time inthe care and rearing of children within the home in an offsetting fashion,increases in the value of time will lead to a shift away from children toless time-intensive activities. To be sure, such a substitution effect may beoffset by a strong income effect, but there are still further grounds tosuppose that both substitution and income effects will tend to lead to anincrease in child quality rather than child numbers.

Earlier in this book, Becker and Lewis explored the consequencesof a simple model of the relation between the quality and quantity of agood (in this case, children) entering the utility function, under theassumption that quality per unit is the same for all units. Increases ineither quantity or quality cause the shadow prices of the other to riseand if, as is plausible, the income elasticity for quality is greater than thatfor quantity, the resulting increase in the shadow price of numbers wouldreduce the apparent income elasticity for quantity, perhaps even tonegative level. Thus, the income effect of the increasing value of humantime should, under these circumstances, lead to a substitution of qualityfor quantity of children, in addition to leading to a substitution awayfrom children altogether to time-intensive goods. This rests of course,on the premise that the productivity of time in rearing children is un-changed or, at least, not greatly increased. Moreover, the pure priceeffects are likely to work in the same direction, since it is plausible thatquantity is more time-intensive than quality of children per unit of equalquality time and that, indeed, the shadow price of quality is likely to bereduced by increases in the quality of the mother's time input (I referagain to Arleen Leibowitz's paper in this volume).

The investments in child quality referred to earlier take two majorforms: (1) sound nutrition and health care, and (2) education, skills, orattitudes conducive to acquisition of further education and skills.

Good nutrition and health care increase youngsters' chances of survivaland may also affect their ability to absorb future investments in intellectualcapital. To the extent that such investments increase the life span, par-ticularly the span of years over which a person can be economicallyactive, such an increase in quality will raise the return to investments inhuman capital which sons and daughters may later wish to make inthemselves. To the extent that better health and nutrition result in areduction in child mortality, they increase the satisfactions accruing toparents from other forms of investment which also raise child quality,for the returns to these investments may then be expected to be enjoyedover a longer period of time on average. Increases in longevity, particularlyof an individual's economically productive years, increase the amount ofhuman time available without increasing population; such an increasewould tend by itself to lower the value of time per unit, but, as we know,

V

POPULATION AND ECONC

most of the effects of beland enhance the qualit'creasing the number olconjecture that better hvestments in human capincrease the returns thet

The second main foreas I have stated,attitudes conducive toMuch investment in hutin the preschool years,time, as remarked, mahuman capital embodiedas the result of theto occur away from thisthat we might expect scould economize on acenters, and the proliferOn the whole, then, Ichildren relative to theicapital which ultimatelunit in the economicall

For reasons which Iare due in part to the piof the family to whichof capital to their chiintangible, human anceconomies, one must asthat which they receiveconomy function inof the motivation,have an effect on therates of return to invemore slowly than, thecapital, parents will behuman capital. Thus,will be intensified by tresulting from theless time-intensive.they ever do—parentscapital and more in thless, as long as investmhuman time will

MARC NERLOVE POPULATION AND ECONOMIC GROWTH 543

tivity of a unit of time inin an offsetting fashion,

It away from children toubstitution effect may bestill further grounds to

ts will tend to lead to an

piored the consequencesuality and quantity of aLity function, under the

all units. Increases inof the other to rise

iality is greater than thatprice of numbers wouldntity, perhaps even toreasing value of humana substitution of qualityto a substitution away

ods. This rests of course,rearing children is un-reover, the pure price

since it is plausible thatildren per unit of equal

of quality is likely to beer's time input (I refer

earlier take two major(2) education, skills, oration and skills.ters' chances of survivalvestments in intellectualease the life span, par-

can be economicallyeturn to investments inlater wish to make ind nutrition result in aatisfactions accruing toiso raise child quality,expected to be enjoyedlongevity, particularly

increase the amount oftion; such an increaseunit, but, as we know,

most of the effects of better health care and nutrition occur in childhoodand enhance the quality of a unit of time in later years more than in-creasing the number of children. On net balance, therefore, I wouldconjecture that better health and nutrition lower the costs of further in-vestments in human capital relative to those in other forms of capital andincrease the returns therefrom.

The second main form which an increase in child quality may take is,as I have stated, through investment in the form of education, skills, orattitudes conducive to later acquisition of further education or skills.Much investment in human capital of this type tends to be time-intensivein the preschool years, although the productivity of a unit of a mother'stime, as remarked, may be especially enhanced by a greater stock ofhuman capital embodied in her, so it is not necessarily true that over time,as the result of the increasing value of human time, substitution will tendto occur away from this form of investment. Nonetheless, it is in this areathat we might expect some induced "technological innovations" whichcould economize on a mother's scarce time. Nursery schools, day-carecenters, and the proliferation of "educational" toys are perhaps examples.On the whole, then, I think we may conclude that increases in quality ofchildren relative to their numbers take the form of investments in humancapital which ultimately have the effect of raising the value of time perunit in the economically active years of adulthood.

For reasons which I feel certain we do not fully understand, but whichare due in part to the presence of children's utilities in the utility functionof the family to which they belong, parents do desire to bequeath a stockof capital to their children. Since the stock of capital, material andintangible, human and nonhuman, is growing per capita in Westerneconomies, one must assume that parents desire to pass along more thanthat which they received from their parents, or that institutions in theeconomy function in such a way as to induce this outcome. Irrespectiveof the motivation, however, the increasing value of human time musthave an effect on the form in which this capital is passed on. As long asrates of return to investments in human capital remain above, or fallmore slowly than, the rates of return to investments in other forms ofcapital, parents will be induced to bequeath a greater part in the form ofhuman capital. Thus, the tendency toward increasing quality of childrenwill be intensified by the bequest motive, despite the opposite tendency,resulting from the increasing cost of time, to invest in bequests which areless time-intensive. But as rates of return tend to equality over time—ifthey ever do—parents should tend to bequeath less in the form of humancapital and more in the form of financial and physical capital. Nonethe-less, as long as investment in human capital occurs, the value of a unit ofhuman time will continue to rise with increases in the stock of capital

544 MARC NERLOVE

per capita, reinforcing the tendency to fewer children of ever-higherquality. Substitution will occur in favor of fewer children of higher qualityand perhaps eventuallyagainst both quality and quantity of children infavor of commodities and knowledge. The "facts" cited at the beginningof this section, suggesting an increasing portion of total capital formationin this century has occurred in the form of human capital, however,indicate that we may be far from the point at which such substitutionbegins to take place against children, quality, and quantity combined.

The outlines of a revised Malthusian model begin to emerge, albeitdimly, from the foregoing conjectures and speculations. In this model,the value of human time and changes in that value over time are pivotal,and the limitations imposed by natural resources are mitigated, if noteliminated, by technological progress and increases in the stock of know-ledge and of capital, both human and nonhuman. The main link betweenhousehold and economy is the value of human time; the increased valueof human time results in fewer children per household, with each childembodying greater investments in human capital which in turn resultin lower mortality and greater productivity in the economically activeyears. Such greater productivity in turn further raises both the value of aunit of time and income in the subsequent generation and enablespersons of that generation to make efficient use of new knowledge and newphysical capital. Eventually, rates of return to investments in physicalcapital, new knowledge, and human capital may begin to equalize, butas long as investment occurs which increases the amount of human capitaLper individual, the value of a unit of human time must continue to in-crease. It is not possible to say whether the diminishing ability of a humanbeing to absorb such investment would eventually stabilize the number ofchildren per household and at what level, given the satisfactions parentsobtain from numbers of children as well as their quality. Nonetheless,over time the model does predict in rough qualitative fashion decliningrates of population growth (perhaps eventually zero rates or even negativerates for a time) and declining rates of infant mortality. These are themain features of the demographic transition.

Much remains to be done if this rough and speculative outline is to betranslated into a true integration of the new home economics with themodern theory of economic growth, particularly as the former evolvesalong more dynamic and empirically relevant lines. First, the model mustbe mathematized—and it is clear that there are many forms in which thismay be accomplished—so that the crucial parameters and behavioralrelations may be isolated. These must then be studied empirically, forwithout quantitative knowledge of the parameters and technologiesinvolved and the key behavioral relations the course of the importantvariables over time cannot be predicted or compared with past behavior.

I POPULATION AND ECON(

The role of the growththose of the econometricthe conceptual foundatirespect to intergeneratunderstand the extentpath, the reasons for seither desirable or necei

I hope this paper rewhere it is we want toought to go from here.

MARC NERLOVE POPULATION AND ECONOMIC GROWTH 545

• children of ever-higherchildren of higher qualityLd quantity of children in

cited at the beginningof total capital formation

capital, however,which such substitution

quantity combined.begin to emerge, albeit

culations. In this model,ue over time are pivotal,

ces are mitigated, if notses in the stock of know-The main link between

ime; the increased valueusehold, with each childtal which in turn resultthe economically activeaises both the value of ageneration and enablesnew knowledge and newinvestments in physical

y begin to equalize, butmount of human capitale must continue to in-

•shing ability of a humany stabilize the number ofthe satisfactions parentsir quality. Nonetheless,itative fashion decliningro rates or even negativeortality. These are the

outline is to beeconomics with the

y as the former evolveses. First, the model mustany forms in which thismeters and behavioral

studied empirically, foreters and technologies

of the importantred with past behavior.

The role of the growth theorist is apparent in the first of these tasks andthose of the econometrician and economic historian in the second. Finally,the conceptual foundations of the new home economics, particularly withrespect to intergenerational transfers, must be clarified if we are tounderstand the extent to which actual growth departs from the optimalpath, the reasons for such departure, and whether collective action iseither desirable or necessary to correct such departures.

I hope this paper represents at least a modest beginning in showingwhere it is we want to get to, if not a set of directions on which way weought to go from here.

V.

COMMENT

The mainof fertility decisions isthat children are goods,time and money—andBut this does not distingation that children tendget us much further. Mcprobably come out asbeginning. In any case,jcosts have a relatively hiThus, any changes thatchanges in total marginrelative to the cost of allis enough. The theory cicompeting ones) of therising cost of time, but ithe major fertility cyclesexplanation yet withinboom of the I 95Os or

My belief is that if wCto put more content intito have children; whatIn terms of the theoretifor shifters of the utilitythe implicit householdfor children rather thaabout children that disgoods. Perhaps we shou"motives" for havingdemand for money ininterdependent motivesprovisions), (2) the proimmortality via one's

The last motive cxchildren (an improvedbasket of consumptionwell-being. The returdeclining secularly, percost-of-time hypothesis.in fertility stems fromchildbearing cohort otheir own parents.

Comment

V

.1

Zvi GrilichesHarvard University

I shall divide my comments into two parts, corresponding to Nerlove'sdiscussion of the "new home economics" and the stylized "facts" thatappear to be building blocks for a future growth model. I am very muchin agreement with Nerlove's exposition of the "new home economics,"but I would like to amplify a bit two of his implied criticisms of the stateof this theory.

One of the major working assumptions of the theory is the existence of acommon family utility function. In his paper in this volume, Gary Beckershows that by introducing the notion of "caring" (or the interdependenceof utilities) one can show that the family will behave as if it has a commonutility function. This will not do, I think, for the analysis of the empiricalphenomena that we are really interested in. What parents care for is notthe utility that their children receive, but the utility function that thechildren have and the resources that they control. Parents care about theconsumption basket of their children; they have preferences over actualactions, not just their subjective outcomes. Much of the within-familyconflict comes from different evaluations of the same consumptionopportunities. For example, since many families appear to subsidize theirchildren's higher education, and since the marginal valuation of familyfunds may not be the same for different family members, it may pay forthe young both to pursue higher education further than appears warrantedon straight rate-of-return calculations (somebody else is paying the cost),and to work less hard at it than the donors of the money would have likedthem to do. In any case, a common utility function cannot explain eitherthe growth of households or their dissolution, or indicate the point atwhich it pays for the young to opt out of it. It is probably also not necessaryfor the analysis of such questions, as is indicated by Becker's model of themarriage market.

546

COMMENT 547 I

rresponding to Nerlove'sthe stylized "facts" thatmodel. I am very much

"new home economics,"led criticisms of the state

heory is the existence of athis volume, Gary Becker(or the interdependence

ave as if it has a commonanalysis of the empiricalt parents care for is nottility function that theParents care about the

preferences over actualch of the within-familythe same consumptionppear to subsidize their

valuation of familyembers, it may pay forthan appears warrantedelse is paying the cost),oney would have liked

n cannot explain eitherr indicate the point atbably also not necessaryy Becker's model of the

The main shortcoming of the "new home economics" for the analysisof fertility decisions is that it assumes too little. The basic postulates arethat children are goods, that all goods are subject to two constraints—time and money—and that children are relatively time-intensive goods.But this does not distinguish children from hi-fl sets! Adding the observ-ation that children tend to use mother's time more than father's does notget us much further. Moreover, given the aims of the theory, this shouldprobably come out as a conclusion, rather than be assumed from thebeginning. In any case, the theory focuses on children as a good whosecosts have a relatively high time component (particularly mother's time).Thus, any changes that have occurred have to be explained in terms ofchanges in total marginal cost (money and time) of this good (children)relative to the cost of all other consumption goods. I do not think that thisis enough. The theory currently provides one explanation (among severalcompeting ones) of the secular decline in fertility, attributing it to therising cost of time, but it fails to provide any convincing explanation forthe major fertility cycles that we have experienced. We have no economicexplanation yet within the framework of this theory either for the babyboom of the 1 95Os or the current rather sharp decline in fertility.

My belief is that if we want to study the demand for children, we haveto put more content into the-theory and start asking why do people wantto have children; what are the returns and not just the costs of this activity?In terms of the theoretical framework used in this volume, I am lookingfor shifters of the utility function, or, alternatively, for factors that changethe implicit household production function. If we are studying the demandfor children rather than for hi-fl sets, we have to ask ourselves what it isabout children that distinguishes them from other time-intensive durablegoods. Perhaps we should go back to some low-level discussions about the"motives" for having children, along the lines of the discussion of thedemand for money in older textbooks. I would distinguish at least threeinterdependent motives: (1) economic security (current labor and old-ageprovisions), (2) the production of reciprocal caring, and (3) an attempt atimmortality via one's offspring.

The last motive explains the attempts to impose an image on thechildren (an improved version of our own) and the interest in the actualbasket of consumption rather than just the summary state (utility) of theirwell-being. The returns in terms of the first two motives have beendeclining secularly, perhaps even faster than the rise in costs implied by thecost-of-time hypothesis. It is my guess that part of the recent sharp declinein fertility stems from the clear recognition on the part of the currentchildbearing cohort of the relatively low rate of return experienced bytheir own parents.

IReferences

Adelman, Irma. "An IA.E.R. 53 (June

Aigner, D. "An Approja Labor Supply Funitems Res. Inst., Univ

Aitchison, John, andCambridge:

Aichian, Armen A.,Costs, and777—95.

Aistrom, C. H. "APsychiatrica et Neurol

Andersen, Ronald, anvices: Social SurveyChicago Press, 1967

Aoki, Hisao. Selected StJapanese.) Tokyo:

"A General ViJapanese.)Jinko Mo

Aoki, Hisao, and Naka1952, 1957, and 19Welfare, 1967.

Arrington, Leonard J.Press, 1958.

Ashenfelter, Orley A.,Ability and Earnin78—86.

Auster, Richard D.;Production of Heal(Fall 1969): 411—36

Bachi, R., and Matramong Jewish Mate40 (April 1962): 20

Bajema, C. J. "EstimSelection in Relationsic Rate of Natural175—87.

Baumol, William J. EYork: Macmillan, ii

Becker, Gary S. "Anand EconomicBureau ConferencPress, 1960.

.548 ZVI GRILICHES

Turning to the last part of Nerlove's paper, I want to take issue withthe two stylized facts that appear to be empirical building blocks for hisgrowth model: the constancy of the rate of return to schooling and apersistent disequilibrium reflected in higher rates of return to the invest-ment in human as compared with physical capital. Both Finis Welchand I have to take some blame for popularizing the first "fact" at a timewhen it was beginning to cease being a "fact." The persistence of relativelyhigh rates of return to schooling through the 1950s and 1960s in the faceof rising schooling levels throughout the economy was probably the resultof an accidental constellation of forces rather than the expression of anunderlying constancy in the economic mechanism. The higher educationboom was sustained by three forces: (1) the first round of cohorts to beeducated after World War II was a relatively small and declining fractionof the population; (2) the post—World War II baby boom increasedgreatly the demand for teachers at all levels; and (3) at about the sametime the government superimposed on all of this a space-defense-research-and-development boom, heavily human capital-intensive, resulting in ascramble for young, educated talent. Unfortunately, these fortuitousinfluences have run their course, The educational system has probablyreached its longer-run equilibrium level, if it has not overshot it. Thespace research and development boom is over now, at a time when thereis, and will be for the next 5 years or so, an annual wave of an additionalone million highly educated workers arriving at the doors of the full-timelabor force. The rates of return to schooling have already started fallingand will probably fall quite a bit further before supply response catchesup with them.

Also, I am not sure that I understand the disequilibrium discussion inNerlove's paper. First, I know of no study that shows that relevant ratesof return, computed in comparable terms, are significantly higher forschooling than for physical investment. The fact that human capital mayhave been growing faster than physical is no evidence for this proposition,and it does not require it.