HUMAN CAPITAL AND ECONOMIC GROWTH Jacob Mincer Working Paper No. 803 NATIONAL BUREAU OF ECONOMIC RESEARCH1050 Massachusetts AvenueCambridge MA… [628505]

NBER WORKING PAPER SERIES
HUMAN CAPITAL AND ECONOMIC GROWTH
Jacob Mincer
Working Paper No. 803
NATIONAL BUREAU OF ECONOMIC RESEARCH1050 Massachusetts AvenueCambridge MA 02138November 1981
The research reported here is part of the NBER's research programin Labor Studies. Any opinions expressed are those of the authorand not those of the National Bureau of Economic Research.

NBER Working Paper 1/803November 1981
Human Capital and Economic Growth
ABSTRACT
Individuals differ in both inherited and acquired abilities, but onlythe latter differ among countries and time periods. Human capital analysisdeals with acquired capabilities which are developed through formal andinformal education at school and at home, and through training, experience,and mobility in the labor market.Just as accumulation of personal human capital produces individualeconomic (income) growth, so do the corresponding social or national aggre-gates. At the national level, human capital can be viewed as a factorof production coordinate with physical capital. This implies that itscontribution to growth is greater the larger the volume of physical capitaland vice versa. The framework of an aggregate production function showsalso that the growth of human capital is both a condition and a consequenceof economic growth.Human capital activities involve not merely the transmission and em—bodiment in people of available knowledge, but also the production of newknowledge which is the source of innovation and of technical change whichpropels all factors of production. This latter function of human capitalgenerates worldwide economic growth regardless of its initial geographiclocus.Contrary to Malthus, economic growth has not been eliminated by popu-lation growth. Indeed, spatial and temporal patterns of the "demographictransition" appear to be congruent with economic growth. Human capital isa link which enters both the causes and effects of these economic—demographicchanges.
Jacob MincerEconomics DepartmentColumbia UniversityInternational Affairs Bldg.— Rat. 809New York, NY10027(212) 280—3676

*HUMANCAPITAL AND ECONOMIC GROWTH
I.IntroductionAs an economic concept human capital is at least two centuries old,but its incorporation into the mainstream of economic analysis and researchis a new and lively development of the past two decades. The need for thisdevelopment became apparent in the 1950's, when the application of empiricaleconomic research to the concerns about economic growth and about incomedistribution revealed major defects not only in our understanding of eachbut also in our way of thinking about these matters. Two types of findings
were especially significant: (1) The observed growth of conventionallymeasured inputs of labor and capital was by far smaller than the growth ofoutput in the U.S. and in other countries for which long time series wereavailable, and (2) Data on personal income distribution, which began to appearwith greater frequency and detail , showedthat the variance of labor incomes,rather than the "functional" differences between returns to labor and tocapital, represented the major component of personal income inequality.The development of human capital theory was a response to these twinchallenges. This response did not require a revolution in economic theoryor a resort to extra—economic explanations which economists sometimes invokewhen answers to pressing questions escape their competence. It merely in-volved the abolition of two simplifying, but as it turned out unduly inhibit-ing assumptions: (1) the restriction of the concept of capital to physical
Prepared for Conference on "Issues in Economic Development,"Mexico City, November 1980.

-2-
capital, even after a more general definition was provided by Irving Fisher,1and (2) the assumption of homogeneous labor which underlies both the conceptof functional income distribution and the measurement of labor input inmanhours.
Fisher's definition of capital as any asset that gives rise to anincome stream requires the inclusion of human capital, even if it cannotbe. bought and sold (it is, of course, rented), and even though investhent.sin such capital often involve non—market activities. But non—market acti-vities are not necessarily extra-economic. To the extent that they involvecosts and returns, whether explicit or implicit, they are amenable toeconomic analysis, even if measurement problems are difficult. The con-tribution of human capital theory to economics does not lie in a reformu-lation of economic theory, but in pushing back the boundaries of economicsbeyond the sphere of market transactions. The payoff is now apparent inboth of the problematic contexts: (1) At the macroeconomic level, thesocial stock of human capital and its growth are central to the processof economic growth. (2) At the niicroeconomic level, differences in indi-vidual human capital stocks and in their growth can explain much of theobserved variation in the wage structure and. in the personal distributionof income.
The application of the human capital concept to economic growth andto labor economics were initially pioneered independently.2 The conceptsare the same, and are applied basically to the same problem: individualeconomic growth at the micro—level, and growth of the economy at the macro—level.

—3—
II.Human Capital and Personal Economic GrowthIndividuals differ in both inherited and acquired abilities, but onlythe latter clearly differ among countries and time periods. Human capitalanalysis deals with acquired capacities which are developed through formaland informal education at school and at home, and through training, exper-ience, and mobility in the labor market. The central idea of human capitaltheory is that whether deliberate or not, these activities involve costsand benefits and can, therefore, be analyzed as economic decisions, privateor public. The costs involve direct expenses and earnings or consumptionforegone by students, by trainees, and by workers engaged in labor mobility.Since production and consumption benefits from these activities accruemainly in the future, and are for the most part quite durable, the costlyacquisition of human capacities is an act of investment. Deterioration ofhealth and erosion or obsolescence of skills represent the depreciation ofhuman capital which is offset, though not indefinitely, by maintenanceactivities such as the production of health and retraining.The general categories of human capital investments can be describedin a life—cycle chronology: resources in child care and child developmentrepresent pre—school investments. These overlap and are followed by invest-ments in formal school education. Investments in labor market mobility,job choice, job training, and work effort occur during the working life,while investments in ha1th and other maintenance activities continue
throughout life.
•School EducationInitially, investment in school education has been the subject of
almost exclusive attention by human capital analysts. While economists

-4-
since Adam Smith recognized the importance of education as a type of pri-vate or social investment, only recently have economists undertakenrigorous conceptual and statistical examination of the evidence on costs,returns, and rates of return to education.The costs of education borne by the student or his parents consistnot merely of tuition and other school expenditures, but also of foregoneearnings——the loss of what the student could have earned if he had spentthe school years in gainful employment instead. Beyond early schooling,foregone earnings are the largest component (over a half) of schoolingcosts.
As in the analysis of physical capital ,thedifference between thediscounted future returns and casts represents the profit or loss on theinvestment. Gains do or ought to induce further schooling and lossesdiscourage it. Another way to represent this decision making process isto calculate that rate of interest which makes the profit equal to zero,that is, it makes the investment just about worthwhile. This rate iscalled the internal rate of return on the investment; further schoolingis encouraged if the internal rate of return on schooling exceeds the rateon alternative investments. The advantage of this approach is that whileindividual discount rates are not observable, internal rates of return canbe calculated given estimates of costs and of earnings streams. Comparisonsof rates of return to education with rates of return on other (say in busi-ness capital) investments, can indicate the desirability of existing alloca-tions or of changes in them, since equality of rates in all types of invest-ments are required for a social optimum.
it is understood, of course, that relevant concepts of costs and bene-fits are real, that is not restricted to pecuniary terms. Education itself

—5—
may be attractive and it may enhancefuture enjoyment of life, apart fromthe monetary gain.Employers pay higher wages to the more educated workers because theirskill and productivity are seen and experienced as greater than that of lesseducated workers. In the absence of strong barriers to supply, thewage differential translates into a r3te of return com-parable to those on alternative human or other investments. Increases indemand favoring more educated workers raise the rate of return on school-ing, inducing growth of enrollments until the increased return has beenreduced back to an equilibrium level. Autonomous increases in supply,given no changes in demand, reduce the rate of return to education and thusbecome self-limiting. The estimated rates of return to schooling in theU.S. have remained relatively stable in the past several decades despitethe continuous growth of educational attainment, suggesting that the trendis mainly a response to the continuous growth of demand for educated labor.If financial and social barriers to education are stronger than inother fields of investment, the rate of return on education exceeds thaton physical capital. Reduction of these barriers brought about by wide-spread growth of family incomes and by public policy has also been a factorin the long—term growth of education. As an example, the growth of educa-tion in the U.S. between 1890 and 1950 was accompanied by a decline in therate of return to education, to levels which no longer exceed the returnto business investment.3
A recent survey of estimates made in 32 countries4 shows that rates
on physical and especially on human capital investments are higher indeveloping countries (LDC's) than in the industrialized (DC's). This is

—6—
-'perhaps not surprising as it reflects the greater scarcity of capital inLX's. More interesting is the finding that, rates of return to human capitalexceed the rates on business capital in LDC's, while if anything the appo-site appears to be true in the DC's. Evidently, the scarcity of humancapital is significantly greater than the scarcity of physical capital inthe LDC's.5
The calculations that are available do not include non—pecuniaryor 'consumptioncomponents" ofcosts or of returns. To the extent thatthese are positive and important in the benefits of schooling (an assump-tion dear to the hearts of educators), the rates are underestimated,though the pattern of their historical changes need not be affected.An important distinction is made between private and social rates.Thus, in calculating private rates, costs and returns to students and theirfamilies are computed from after-tax data, and schooling costs do notinclude public financing of schools. In contrast, the calculation ofsocial cost is based on before—tax- earnings, and school costs are totalcosts of the relevant school system (per student) regardless of the sourceof financing. The real difficulty in calculating social rates of returnis the problem of measuring externalities. To the extent that the gainto society exceeds the sum of gains to students, social returns are under-estimated. An assumption of public policy which is difficult to verifyand.to quantify is that such externalities are substantial and positive.It is often suggested that these externalities include, among others,informed and responsible citizenship, comunications skills, lawful behavior,and standards of health. The existence of such externalities is invoked tojustify public efforts to stimulate educational investments. Such efforts

—7—
can take the form of a publicly owned school system and/or directsubsidies to students. The extent of required support is alwaysdebatable, as the magnitude of externalities is unknown.There are also other reasons for public intervention. This isthe concern with the distribution rather than with the total volumeof educational investnents. Helping children of the poor to acquirea minimal degree of earning power is an objective for which schoolingis also viewed as an instrument. Since poverty is often viewed asa relative concept, thefl amount of minimal universal governmentsupported education has been progressively lengthening as averageeducation (and income) have increased. It is not always clear, how-ever, to what extent these policies are efficient in alleviatingpoverty. There is some evidence6 that public spending on primaryeducation tends to be redistributive toward the poor, but that abovethat in LDC's, and above secondary education in DC's, the oppositeis likely to be true, on balance, since children of the poor areless likely to acquire higher education.
2. Post—school Human Capital Investments7There is no reason to believe that human capital investments ceasewith the termination of schooling. The educated have higher earnings, butthe earnings are not fixed. They grow over the working life, albeit at adecelerating pace. This growth is additional to, and largely independentof economy—wide trends in earnings. These patterns of growth also differamong persons whose education is similar.
The economic interpretation of lifetime earnings growth is as fol-lows: Wages of a worker are proportionate to the size of his human capital

–
stock.Thus, wage differentials among workers are due primarily to dif-ferencesin human capital stacks, not in the "rental price" employerspay per unit of the stock. The individual P humancapital stock growsover the life cycleS by means of investment, which is initially in school-ing, later in job choice, job training, job mobility, and in health. Atany stage, the level of earnings depends on the size and utilization ofthe human capital which accumulated up to this point, and its growth dependson the rate of net additions to the stock, that is, on the net investmentrate. The deceleration in the rate of growth which is observed in indi-vidual earnings reflects the rate of decline of investments as the workerages. Investments diminish over time because (1) benefits decline as the
payoff period (remaining work life) shortens, and (2) the opportunity costsof time, which is an input in the learning process, are likely to rise overthe working life. While gross investment proceeds at a slackening ratethroughout working life, net investments (gross minus depreciation) vanishor turn negative earlier. This happens when depreciation (including obso-lescence) begins to outstrip maintenance, a progression which eventuallybrings about retirement.An alternative interpretation of the earnings profile is that it isan intrinsic age phenomenon; initial productivity growth corresponds toinherent biological and psychological maturation, while later stabilityand decline are due tO first stable then declining physical and intellec-
tual vigor.In the perspective of human capital, this view is incompletesince it explains the earnings profile solely by a life—cycle pattern ofthe depreciation rate, seen as negative in early years, zero in middlelife, and positive in later years. There is evidence, however, which

—9—.
indicates that this inherent age—depreciation factor affects earnings onlyto a minor degree, except at teenage and in the near or post-retirementyears; in data where age and length of work experience are statistical lyseparable, levels and shapes of earnings curves are mainly a function ofexperience rather than of age. Moreover, earnings profiles differ byoccupation, sex, and other characteristics in systematic ways that cannotbe attributed to agtng.One may also interpret the shape of the earnings profile as a"learning curve," or a reflection of growth of skills with age and exper-ience known as "learning by doing." This view is not at all inconsistentwith the human capital investment interpretation, as long as opportunitiesfor learning are not costless. Since more learning, hence a more steeplyrising wage is available in same jobs compared to others, qualifiedworkers would gravitate to such jobs if learning were thought to becostless. In consequence, entry wage levels in such jobs would be re-duced relative to entry wages elsewhere, for workers of the same quality,thereby creating opportunity investment costs in moving to such jobs.Thus, it is not merely training on the job (formal or informal), but alsothe processes of occupational choice that give rise to investments beyondschooling. Similarly, geographic mobility and other labor turnover insearch of higher real earnings represent investments in human capitalIt follows that barriers to occupational choice and to jobmobility reduce the opportunities for investment in human capital. Theelimination of such barriers increases individual economic growth andthe overall efficiency of allocation of resources in the economy,hence total product.

—10—
Empirical economic research indicates that the relation betweenschooling and post—school investment is positive: More educated peopleinvest more in the labor market. One interpretation is that ability andopportunity factors which induce individuals to have more schooling affecttheir post—school behavior similarly, even though the correlation is farfrom strict: Abilities and opportunities change over the life—cycle,and there is a fair amount of substitution between the two forms of skillaccumulation. Another interpretation is that schooling improves the effi-ciency with which people can absorb learning on the jab, leading therebyto greater job investments. This hypothesis is consistent, in a dynamiccontext, with evidence on the so—called "worker allocative effect" pro-pounded by Schultz (1976) and Welch (l97a). Their proposition that educa-tion promotes the adjustment to technological change has been documented,mainly in studies of agricultural production activities. The more limitedmacro-economic evidence of a positive relation between rates of return toschooling and.rates of economic growth is also suggestive.9
3. Preschool Investments and Women's EducationInherited abilities, or what is called the 'original" endowment isan important part of the human capital stock, yet the line between heredityand Environment is by rio means clear. Much of the physical and intellec-tual deficiency shown by infants born in poor conditions can be avoidedby improved nutrition of mothers and sanitary environments for childbirth.Similarly, subsequent child care represents an investment in better adulthealth and so in greater productivity of the adult worker.Especially In low income countries, the effects of a healthier childrearing on adult productivity are double: Not only is a healthier adult

—11—
more productive but he also lives longer10 Consequently, the incentivesto invest in lengthier schooling and training increase, since with thelengthened payoff period, the profitability of such investments increases.Thus, it is inappropriate to view reductions in mortality with alarm as acause of the "population explosion." The mitigating effects on populationgrowth and improvements in work quality eventually predominate, sincethe costs of investing in child quality, including health and education,represent a powerful force toward reduction of family size; given thefamilies' limited resources. Indeed, research has shown that even at thesame level of family income, children in smaller families tend to behealthier, more intelligent and better educated.Much of the accumulation of a person's human capital takes Ølacein the home, particularly during the pre—school stage of the life—cycle.It appears that education of parents is a significant influence in thisprocess, even after controlling for family income and numbers of siblings.This suggests that aside from expenditures on schooling and health,child care is also an important qualitative input into the productionof human capital. The time inputs are usually those of themothers who take the major child care responsibilities and reduce theirmarket activities to engage in them. The consequent reduction in theirearnings may be viewed as a partial measure of opportunity costs of theseinvestments. So viewed, the opportunity cost o ehild care is greater, formore educated women. The observed positive effects on children's health,intelligence, education and future earning power may thus be viewed as anindirect return on the investment in maternal education.An important consequence of the larger opportunity cost per unitof

—12—
time spent in child care by more educated mothers is the reduction of totaltime so spent. This is accomplished largely by•a reduction in the numberof children. The strong inverse relation between fertility and educationof mothers has been documented repeatedly. Thus the growth of women1s edu-cation and of their wages induces declines in fertility coupled with in-creased investments in the resultant smaller number of children per family.Since, in most countries, evsi educated women spend less time in thelabor market than men, the direct earnings benefits of education are smallerfor women. From this point of view, It might seem that the provision ofequal amounts of education to both sexes is wasteful . However,if bettereducated mothers produce greater human capital In children and a betterquality of family life, apart from contributing to family money income,educational equality need not be questioned. Indeed, it is rarely questionedas a matter of public policy.It appears, however, that private schooling decisions are still verymuch influenced by the expected participation in the labor market, and
therefore by the directly expected payoff in earnings. In the U.S. sexdifferentials in enrollment now appear only at the postgraduate universitya..-(r.. o.4 &jclevel .IxutatinAmerica there remains a pronounced differential aboveprimary school enrollment.

—13—
III. Human Capital and National Economic Growth1. Human Capital as a Factor of Production. The micro—economicanalysis of investilent in human capital is the underpinning of our under-standing of the contribution of human capital to the aggregate level ofincome and to its rate of growth. The micro—economic view is most directlyapplicable to the analysis of labor heterogeneity and of the resultingwage structure. Given sufficient labor mobility, wages tend to be simi-lar for the same human capital stock in various employments, regardlessof differences in size and quality of other factors of production insuch employments. Equilibrium wage differentials within the economy maytherefore be viewed as reflecting solely differences in individual magni-tudes of human capital stocks. Although international mobility of laboris not negligible and it mitigates somewhat the disparities in wages of thesame human capital in different countries, national wage levels' differ becauseof differences in volumes of human capital as well as of other forms ofcapital. For the understanding of macro—economic differences in levelsand in growth of income It is best to start with the view of human capitalas a factor of production alongside physical capital in an aggregateproduction functionThe traditional trinity of factors of production contained landviewed as fixed, "original and indestructible," labor measured in numbersand hours, and capital •restricted to tangible plant and equipment. It isnow well recognized that this conception is false. The notion of a quantityof land as a fixed factor of production has already been discarded prior tothe realization that the measurement of labor in manhours is entirely inade-quate. As T. W. Schultz has emphasizedl2differerices in amount and "original
quality" of arable land (in terms of land—population ratios) do not at all

-14-
neip in accounting for differences in income levels among countries. Exper-ience and research have shown that it is not the quantity and the originalendowment of land, so much as the improvement or modernization of agri-culture that matters. Inferior raw lands and even deserts have been trans-formed into superior productive resources, while total acreage declined.Investment in modernization of agriculture is a capital investment, andthe capital nature of land is now fully recognized.The capital nature of the sources of labor services is now also re-ceiving its proper recognition. The inadequacy of the traditional viewof labor in the field of growth accounting is well known. But the biaseswent beyond description to affect policy: The misunderstanding of thenature of expenditures on health, education, labor mobility, and informa-tion as consumption which reduces saving lead to investments in steelmills rather than in people.Land, by itself, is no longer a limiting or critical factor. Butthe quality and behavior of people is increasingly recognized as such.Indeed, it appears that indexes of human capital, such as average levelsof education, are more strongly correlated with average income levelsacross countries than measures of physical capital per unit of labor. 13
Although suggestive, this finding is not conclusive since the demand foreducation as a consumer good is income elastic. In this sense, educationis an effect rather than a cause of income. The role of education as acause, however, is evident from the micro—economic findings that the rela-tion between education of persons and their own future income is strongand largely unaffected by parental income, even though parents' incomedoes affect the amount of education their children receive.14

—15—
Some critics question the inference that education increases pro-ductivity from the observation that it increases wages, and still othersassert that schools do not affect skills but serve merely as a filter tosort differences in ability which exist independently of schooling. Ifso the micro—economic relation between education and income would notcarry over to the economy as a whole. This argument is contradicted byresearch: Studies of empirical production functions have shown that notonly differences in wage rates but differences in productivity are relatedto differences in education and training of the labor force across states,regions, and over time.15This is not to say that the screening or sorting function of educa-tion is unimportant or unproductive. Indeed, the search for talent bythe school and by the student are activities no less productive than thesearch for any other scarce natural resource. Human capital is augmentedboth by learning and by selection. The interaction of the two is efficient:The more able student learns more at the same cost.The view of human capital as a factor of production coordinate withphysical capital implies that its contribution to growth is greater thelarger the volume of physical capital. This relation is symmetric: Thecontribution of physical capital is larger the higher the average level ofhuman capital. In this light, the success of the Marshall Plan in Europeand the failure of foreign aid to LDC's are perhaps not surprising. Toquote Harry Johnson: "Europe had available the industrial and comercialorganization, and the skilled people required for modern industry; what itlacked was precisely physical capital which was largely destroyed or obsolete.The problem of LDC's was different: They lacked virtually everything

—16—
necessary for a higher standard of economic productivity, and the injectionof only one element (physical capital).was found to be both wasteful anddisappointing."16For the more recent period the problem of absorption of massiveamounts of physical capital in the human capital—poor OPEC countries isanother example of the significance of compiementarity between the twoforms of capital. But, while physical plant and equipment can be acquiredor built quite rapidly, the development of a significant and broadly basedlevel of human capital of a nation is a lengthy process which involvesprofound social and cultural changes)7The framework of an aggregate production function makes it clearthat the growth of human capital is both a condition and a consequence ofeconomic growth. The growth of human capital raises the marginal productof physical capital which induces further accumulation of physical capitalthus raising total output both directly and indirectly. Conversely andsymmetrically, the growth of physical capital raises the marginal productof human capital. This produces an increased demand for human capitalrelative to unskilled labor, if human capital is more complementary withphysical capital than is unskilled labor)8 The resulting increase in theskill wage differential exceeds the increase in (opportunity) costs, sothe acquisition of human capital by students and workers becomesrnore pro-fitable. As already indicated, the continuous long—term growth of humancapital in the U.S. and elsewhere is consistent with this interpretation of
supply responses to growing demand.The differential shifts in demand forskilled and unskilled laborimplied by the complementarity hypothesis also tend to produce the well known

—17-
skill differentials in unemployment rates, observable in most countrieswhich experience economic growth. The greater cyclical stability ofemployment of skilled labor is also consistent with the hypothesis that -skilledlabor is complementary with fixed plant and equipment. Actually,recent research suggests that both wages and employment of skilled laborare relatively insensitive to the business cycle, because human capitalacquired on the job contains elements of finn specificity which makeseparations unprofitable to both workers and firms)9Growth of human capital is also spurred on the supply side bygrowth of family incomes. Since markets for financing of human capitalinvestJnts do not exist, growth of income enables larger numbers ofpeople to self-finance their human capital investments. In poor countries,these financial restrictions create monopolistic advantages for the childrenof the wealthy, and high rates of return on human capital. Both are reducedby the spread of education made possible by growing incomes. However,human capital growth due to growth of family incomes is eventually self—limiting, when rates of return become sufficiently depressed in consequenceof "overeducation.' Public subsidies are also self-limiting in the samesense, and they may become unprofitable from a social point of view (whenthe social rate of return on human capital drops below the correspondingrate on physical capital) before they inhibit private incentives. It fol-lows that for a sustained growth of human capital we must look to increasingmarket demands for skills and technology.

-18-
2. HumanCapitaland Technology1 Although the effects of humancapital growth and some of its causes can be described in the framework ofan aggregate production function in which technology is fixed, few will argueagainst the view that growth of technology Is the ultimate force which propelsall factors of production by increasing their productivity. A fixed tech-nology may be maintained for analytical convenience by viewing all technicalchange as embodied in human and in physical capital.2° Whether or not sucha device is purely semantic, I think it is helpful to distinguish betweenthe stock of human capital as a standard factor of production and the stockof knowledge as the source of technology. Human capital activities involvenot merely the transmission and embodiment of available knowledge in people,but also the production of new knowledge which is the source of innovationand of technical change. Without new knowledge, it is doubtful that largerquantities of existing physical capital, more widespread education and health wouldcreate a continous growth in productivity on a global scale. In a fundamentalsense, modern economic growth is a result of the scientific revolution,that is, of the growth of systematized scientific knowledge.The geographic origin and spread of the industrial revolution sincethe 18th Century supports this view and the pivotal role of human capitalin generating and facilitating it. The industrial revolution started withthe scientific revolution in the Northwest of Europe and spread most rapidlyto those areas where educational development has made the transfer of tech-nology most feasible.It is clear now that the process of growth and diffusion is worldwide.Human capital as embodiment of skills is a convenient conceptualization ofits role as coordinate factor of production in its contribution to nationaleconomic growth. Human capital as a source of new knowledge shifts production

—19—
functions upward and generates woriwide economic growth.Even though "knowledge knows no boundaries," its utilization requireslocal adaptation which is more costly the more dissimilar ("distant") theeconomies and societies to which it is transmitted. Moreover, as technicalprogress continues, the slower the diffusion the wider the technologicalgap between the initiators and the "latecomers." Consequently, the capacityto absorb and to adapt new technology requires an increasingly specializedandsophisticated labor force backed by a broadly educated population.For reasons that certainly make sense in the technology exporting countries,the imported modern technology is capital and skill intensive. Thus, problemsof "labor absorption" are added to the difficul ties of absorbing moderntechnology.21Yet, the disadvantages of factor bias are transcended by the advan-tagesof being able to skip several generations of technology in a shorttime. At any rate, it is only the most modern technology that is trulyavailable. Older vintages which may be more labor intensive are not usablewithout complementary or ancillary industries which are obsolete. Even ifthe initial effects on the creation of highly productive employment arerelatively small, the simultaneous adaptation of human capital by jobtraining and some job redesign can help to widen the process. Initially,the pressure of modernization is most acutely felt at the highest educationlevels: specialized scientists, technicians and researchers are needed toadopt, master, and modify the new technologies. But only widespread educa-tional growth, especially at basic levels of literacy and numeracy can leadfrom islands of modernity to a complete transformation of the economy.22

-20-
111.Human Capital and Population
According to Malthus, economic growth can only be sporadic: It isself-defeating, since it produces population growth which in turn swallowsall the gains. This theory has long been contradicted by empirical evidence.The notion that this hypothesis may be applicable to LDC's which was enter-tained by some is also being discredited by events. Economic growth has notbeen eliminated by rapid population growth in these countries.Moreover, the patterns of population changeassociated with the "demographic transition' in the West are now beingvisibly repeated in the rest of the world?3 Indeed, the congruence ofspatial and temporal patterns of economic growth and demographic changesuggest an important interaction between the two. Human capital is a linkwhich enters both the causes and effects of economic-demographic changes.Human capital, or population quality, was left out of Malthusiantheory. The theory actually omits any economic motivation. It presentsa strictly biological view of mortality as a mechanism which adjustsnumbers of people to available resources. The contrary facts of economicgrowth and of the demographic transition have led to a reformulation ofpopulation theory in terms of parental decisions about numbers and"quality of children."24 In primitive, premodern regimes of very highmortality, especially in an agricultural setting, unlimited fertility maybe viewed as a rational- response, which is also (or therefore?) culturallysanctioned. Declines in mortality, brought about by publichealth measuresor by higher levels of living bring about the need for family size decisions,given the family's limited resources. Implicitly such decisions must con-sider both material and 'psychic' costs and returns from children. Intentio

-21—
about human capital formation in children, or child 'quality" play a partin the decision. Given the family budget, resources spent on "quality"compete with the number of children the family might otherwise want.This trade—off becomes pronounced in the context of economic growthwhich raises the payoff to human capital formation.In the West, mortality reductions initially resulted in increasedfertility but after a long lag, they were followed by fertility declines.Surviving average family size grew initially, but eventually declined tothe present-day low levels. Roughly speaking, family size begins to de-cline when fertility rates drop more sharply than mortality rates.Although even exogenous declines in mortality tend to induce declines infertility, it appears that for birth rates to fall more than death rates,the additional stimuli of economic growth and of widespread education are
necessary.This generalization is supported both by the history of DC's andby current experience inLDC's. An intercountry analysis of changes duringthe past decade (1965—1975) in Latin America25 showed that declines inbirth rates were positively related to declines in death rates, but thedeclines in births were steeper than the declines in deaths only in coun-tries whose rates were above average during the decade and educationalenrollments of the population aged 5-14 were significantly above averageat the outset. The regression analysis shows that at a rate of 2% growthof per capita income the enrollment rate must be at least 80%, to gen—erate a reduction in family size. With a growth rate of 3%, the minimumenrollment rate is 60%.

-22-
Butwhat is there in the process of economic and educational growth thatmakesincipientincentives toward reductions in family size widespread, effective,and progressive? In a way, the emergence of strong growth implies that someof the cultural inhibitions to rationalism have already weakened. More directly,economic theory contains three implications of economic growth which point todeliberate reductions in family size: (1) urbanization, (2) the rising costof time, and (3) educational growth:(1) Since demands for agricultural products are relatively income andprice inelastic, the growth of productivity reduces the demand for farm laboy,which in turn flocks to cities in search of employment and higher wages. Withchildren less productive and more costly to raise in the city than on the farm,incentives of migrants to limit family size are strong.(2) The growth of wages in the labor market attracts people from non—market activities (households and subsistence sectors) to the labor market. Tqthe extent that child—rearing is a time—intensive activity, increases in marketwages represent a rising foregone cost of time spent in child care rather thanin gainful work. Therefore, incentives of women to limit family size and toenter (or stay in) the labor market appear and grow. This is especially trueof educated women, since opportunity costs increase with education. A strongnegative correlation between education of mothers and family size has beenwidely documented.26(3) With growing incomes and industrial demands for literate, dis-ciplined, and skilled labor, both private and public demands for educationincrease. At the family level , thedemands for prolonged education ofchildren represent an additional incentive to substitute "quality" forthe quantity of children, as the reduction in numbers of children increases

-23—
the available family resources, per child. The inducement•to invest inquality and in greater future earning capacity_of the children is strengthenedby increased life expectancy as it constitutes a lengthened "pay—off period'on the investments. In turn, when the educated children become parents, theytend to have more favorable attitudes and more infonnation about birth controlbehavior and greater demands for education and health of their children.In sum, we should expect growing urbanization, education, female laborforce participation, and declining family size to follow economic growth.Such trends are, indeed, widely observed under conditions of sustainedeconomic growth, although intensities and time lags in these processes canand do differ from one setting to the next. For example, growing marketwages may induce women into the labor market without reducing their fer-tility, if the extended family and cheap domestic service can help inchild rearing, and if the nature of work, such as farming or cottage indus-try are not incompatible with the immediate presence of children. Also, fora time, growth of wages may reduce fertility without increasing the laborforce: This happens when women employed in the occupations just describedincluding domestic service move to better paid factory work. All the same,the extended family institution and the occupations compatible with uninter-rupted mother's child care eventually decline as incomes and education con-tinue to grow, and all the predicted effects become apparent as they do inthe industrially developed countries.The significance of these demographic events for the quality oflabor is twofold:(1) High birth rates imply an age distribution of the population
heavily weighted toward youth. For example close to one—half of the Mexican

—24—
population is less than 15 years of age. This represents a heavy burden onthe economy, since the consumption and educational needs of the young popu-lation are paramount, and their economic contribution small . Continuationof declines in birth rates will change the age distributiontoward a more productive labor supply.
(2) Beyond improving he quality of the labor force via changes inage distribution, reductions in the size of large families apparently alsoaffect educational progress, as was already alluded. Families with fewerchildren can more readily afford educational expenditures. If the frequencyof large families is greater among the poor, the induced demographic changeshave important positive effects on the future distribution of income and onsocial mobility.
In this brief exposition it was not possible to do more than sketchthe theory and allude to some of the empirical research which documents thevital and manifold roleof human capital formation in personal, national,and global economic development. I think it is fair to conclude that evenif substantial levels of human capital may not be a prerequisite for an accelera-tion of economic growth at a certain time and place, the concurrent growth anddiffusion of human capital appear to be necessary to insure sustained economic
development.

—25—
Footnotes
1.Irving Fisher, The Theory of Interest, Macmillan, New York, 1930.2. For an exposition of these origins see Paul J. McNulty, TheOrigins and Development of Labor Economics, pp. 192—200, MIT 'Press, 1980.The works cited are: Jacob Mincer, "Investment in Human Capital andPersonal Income Oistribution,nu Journal of Political Economy, August 1958,Theodore w.Schultz,"Investment in Human Capital,' American EconomicReview, March 1961, and Gary S. Becker, Human Capital, Columbia UniversityPress, 1964.3. Findings in Becker, p.cit.4. George Psacharopoulos, Returns to Education, Elsevier, 1973.Table 5.3, p. 86 shows that in a sample of countries whose per capitaincome was under $1,000 in the early 1960's, the average rate of returnon education was 19.9% and on physical capital 15.1%. In countries withhigher income, the educational rate of return was 8.3% while the rate onphysical capital was 10.5%. An estimate for Mexico at about that timewas 21.9% for education and 14.0% for business capital. These estimatesappear in Martin Carnoy "Rates of Return to Schooling in Latin America,"The Journal of Human Resources, Summer 1967.
5.Indeed, ir a recent meticulous study, "Perspectives on Capitaland Technology in Less—Developed Countries" (1978), Arnold Harbergershows that around 1970 rates of return (social or private) on physicalcapital no longer differed between advanced and less developed countries.One may speculate that the growth of international investments has ledto this result. Does this mean that physical capital is no longer scarce

—26—
in developing countries? Not at all, its quantity is still limited bythe scarcity of human capital: An increase in the latter will raisethe rate of return on the fonner, increasing the demand for it untilmarginal returns on both fonns of capital have equalized.6. See World Bank, World Economic Report, 1980, pp. 49—50.7. This section draws on my "Human Capital and Earnings," inEconomic Dimensions of Education, a Report of the National Academy ofEducation, May 1979.8. Theodore W. Schultz, "The Value of the Ability to Deal WithDisequilibria," Journal of Economic Literature, September 1975. FinisWelch, "Education in Production," Journal of Political Economy, Febru-
ary 1970.9 Carnoy,cit.,and Psacharopoulos, 9R cit.10 For an empirical study of India, see Rati Ram and TheodoreW. Schultz, "Life Span, Health, Savings, and Productivity," EconomicDevelopment and Cultural Change, April 1979.
11.See World Bank, 2E• cit., p. 47.12. In his Nobel Lecture, "The Economics of Being Poor," publishedin Journal of Political Economy, August 1980.13. Anne 0. Krueger, "Factor Endowments and Per Capita IncomeDifferences," Economic Journal, September 1968. Also Psacharopoulos,op. cit.
14.For a survey of U.S. findings see Mincer,op. cit. Similarfind-ings are shown by Carnoy cit. for Latin America. However, his datacontain father's occupation, rather than parental income.

—27-
15. For references, see Mincer, 2R cit. Also, World Bank,cit., p. 38, and Victor J. Elias, "Sources of Economic Growth in LatinAmerican Countries," Review of Economics and Statistics, August 1978.16. Harry G. Johnson, On Economics and Society, University ofChicago Press, 1975, p. 283. Japan should be added to the Europeanexample in this quotation.17. See Lawrence Stone, ed., Schooling and Society, Johns HopkinsUniversity Press, 1976; and John C. Caldwell, "Mass Education as aDetenninant of Fertility Decline," Population and Development Review,June 1980.18. Some evidence is provided in Zvi Griliches, "Capital-SkillComplementarity," Review of Economics and Statistics, November 1969.Complementarity of human capital with technology would produce thesameresults.This is stressed in Schultz (1975), 22.• cit., and Welch,22. cit.Evidence from time series is provided in Manoucher Parviri,"Technological Adaptation and Income Growth," Ph.D. Thesis, ColumbiaUniversity, 1963.19. See Becker, .cit.Also, Jacob Mincer and Boyan Jovanovic,"Labor Mobility and Wages," in Sherwin Rosen, ed., Studies in LaborMarkets, University of Chicago Press, 1981.20. See Zvi Grilicles and Dale W. Jorgenson, "The Explanation ofProductivity Change," Review of Economic Studies, July 1967.
21 .Theproblem of "labor absorption" has apparently been over-stated in the literature on economic development. According to Harber-ger (1978), "it is not true that the less-developed countries are

-28-
condemned by modern technology to use (in their modern sectors) capitaland labor in proportions that are very similar to those employed in theadvanced economies.' His own estimates show that modern sectors areseveral—fold more labor intensive (per unit of capital) in LUG's thanin advanced economies.22. See Anderson and Bowman in Stone, 22.• cit. Also, J.C. Caldwell,2E cit.23. See Figure 5.3 in World Bank, op. cit., p.65'.S-
24.See the compendium edited by T. W. Schultz, Economics of theFamily, University of Chicago Press, 1975. Also, World Bank, 9!- itt-'Caidwell,pp. cit., and the UBER Conference edited by Richard Easterlin,Population and Economic Change in Developing Countries, University ofChicago Press, 1976.25. Jacob Mincer, "PopulaçAo e Força De Trabaiho," in RevistaBrasileira Da Economia, December 1975.26. See Schultz, Easterlin, Caidwell, j. cit.