The Production of Human Capital and the Life Cycle of Earnings

Author(s): Yoram Ben-Porath

Source: The Journal of Political Economy, Vol. 75, No. 4, Part 1 (Aug., 1967), pp. 352-365
Published by: The University of Chicago Press
Stable URL: http://www.jstor.org/stable/1828596
Accessed: 11/05/2010 03:30

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=ucpress.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to The
Journal of Political Economy.

http://www.jstor.org
 THE PRODUCTION OF HUMAN CAPITAL AND
THE LIFE CYCLE OF EARNINGS
YORAM BEN-PORATH*
University of Chicago

T HE application of capital theory to ceive returns on their investment. The

decisions on individual improve- purpose of this paper is to combine that
ment, and in particular improve- part of the argument concerning the
ment of earning capacity, has provided a demand for human capital with a more
framework for the understanding of explicit treatment of the supply, or cost
many aspects of observed behavior re- conditions, facing the individual.
garding education, health, occupational It is hard to think of forms of human
choice, mobility, etc., as rational invest- capital that the individual can acquire as
ment of present resources for the purpose final goods-he has to participate in the
of enjoying future returns. The formula- creation of his human capital. His own
tion by Friedman and Kuznets (1945) abilities, innate or acquired, the quality
and the significant development of the of co-operating inputs, the constraints
theory by Becker (1962, 1964) and and opportunities offered by the insti-
Mincer (1958, 1962) provided a novel tutional setup-all determine the "tech-
view of the life cycle of earnings by nology," or the production function.' To-
linking it to the time profile of invest- gether with the relevant factor prices,
ment in human capital: People make the properties of the production function
most of their investments in themselves determine the optimal way in which any
when they are young, and to a large quantity of human capital is to be pro-
extent by foregoing current earnings. duced and determine the cost of produc-
Observed earnings are therefore relative- tion. I shall show how the production
ly low at early years, and they rise as function (through supply or cost condi-
investment declines and as returns on tions) enters into the determination of
past investments are realized. The main the optimal path of investment, analyze
reason why investment is undertaken some of the implications for the individ-
mostly by the young is that they have a ual's allocation of time, and demonstrate
longer period over which they can re- how the life cycle of earnings can be
* This paper is based on part of my Ph.D. dis- affected by various properties of the
sertation submitted to the Department of Eco- production function.
nomics, Harvard University, February, 1967. I The basic model generates some of the
benefited from discussing earlier drafts with my qualitative characteristics of the ob-
adviser, Professor Simon Kuznets, and with Profes-
sor Rodney A. Dobell, Gur Ofer, and Christopher A. served life cycle of earnings-typically,
Sims; I received many useful comments from Profes- an initial period of no earnings followed
sors Gary S. Becker, Zvi Griliches, Theodore W.
Schultz, Lester G. Telser, and other friends. Finan-
by a period in which earnings rise at a
cial support was provided by a Ford Dissertation declining rate and, eventually, decline.
Fellowship and a grant from the Ford Foundation
to the University of Chicago for Research on Invest- 1 Such a production has been recently introduced
ment in Human Capital. also by Becker (1966).
352
 HUMAN CAPITAL AND THE LIFE CYCLE OF EARNINGS 353

Actual, or observed, earnings turn out to maximize the discounted value of any
be always lower, to change faster, and to time t of disposable earnings (defined
peak at a later age than the attainable below) from t to T, where T is assumed
maximum, the earning capacity of the with certainty to be the end of life.
individual. The reverse is true of the b) Given the optimal time path of dis-
relation between observed earnings and posable earnings, the individual decides
earnings net of all investment costs: the on the timing of the consumption. This
former are always higher, change slower, is the point of departure of the life-cycle
and peak at an earlier date than the theories of consumption, which take the
latter. The model thus specifies the stream of earnings as given and explain
nature of the bias that may exist when consumption as dependent upon it.
earnings are used, as they often have The stock of human capital is defined
been, to infer changes in productive as a concept analogous to "machines" in
capacity with age. the case of tangible capital. There is a
market in which the services of human
I. THE MODEL
capital are traded, and a rental, ao, is
In addition to the assumptions incor- determined for the services of a unit of
porated in the production function, to human capital, K, per unit of time. The
be discussed later, I assume: sum of the services offered in the market
1. Individual utility is not a function by various individuals is an input into
of activities involving time as an input. the production of other goods and serv-
2. There is a fixed amount of time to ices. It may well have a diminishing mar-
be allocated every period to activities ginal productivity, which will cause
that produce earnings and additions to a downward-sloping aggregate demand
the stock of human capital. curve for the services of human capital.
3. The stock of human capital, K, of Each individual is assumed, however,
which every individual has some initial to possess only a small fraction of the
endowment, is homogeneous and subject total homogeneous stock of human cap-
to an exogenously given rate of deterio- ital in the economy and is regarded as a
ration, 3. perfect competitor facing a given rental
4. The stock of human capital is not ao, which is independent of the volume
an argument in the individual's utility of services that he offers in the market.
function. Earning capacity at time t, Yt, is there-
5. Unlimited borrowing and lending fore the maximum services of human
take place at a constant rate of interest, capital the individual can offer in the
r. market valued by the rental ao.
The first two assumptions express the
Yt = aoKt.(1
fact that leisure is ignored in this analy-
sis. In conjunction with the other as- Let Et be disposable earnings in period
sumptions they allow the partition of it-the portion of current earnings dis-
the individual's decision-making into posable for purposes of consumption or
two stages: the purchase of non-human assets. It
a) The individual allocates the given may be smaller than earning capacity if
periods of time between earning and pro- the individual engages in production of
ducing human capital and finds the cor- human capital; other uses for time are
responding outlays on investment that excluded by assumption. The difference
354 YORAM BEN-PORATH

(Y, - Et) is the cost of investment It; (2) and ignoring (3), we get (6) as a
it depends on the production function condition for a minimum of (5):
and on input prices. aosKt 1 6
Let (2) be the production function of PdDt f2(
human capital:
(2) From (2) and (6) substituted into (5) we
Qt= 0( sKt)P1DO2,
get investment costs as function of out-
where Al, 12 > 0 an d 31 + /2 < 1; Q is
put:
the flow of human capital produced. D is f1 + 032 (lPd\ 2'(01+2)
the quantity of purchased inputs, the it=
t ao? )
01 #2a0 (7)
price of which is denoted by Pd; St is the
fraction of the available stock of human
capital allocated to the production of
human capital, so that stKt is the quan- The objective of the individual at any
tity of human capital allocated for the time t is to maximize the present value
purpose. If activities are not "mixed," of his disposable earnings:
that is, if there is no joint production of T

earnings and of human capital, then st Wt =| e-rv[aoK(v)-I (v) ]dv. (8)

is also the proportion of time devoted
to the production of human capital. The The objective expressed by (8) is treated
fraction st is constrained by the condition as applicable to an individual from birth;
up to a given age actual decisions are
O < st:< 1 . (3) made by his parents, and if they were to
(The properties of the production func- take into account the full future life of
tion including the assumption of de- their child, (8) still expresses the relevant
creasing returns to scale are discussed in maximand under the assumptions of the
a subsequent section.) The rate of change model. If the parents do not take into
of the capital stock is given by (4): account the full economic life of their
child, or in the extreme case where only
Kt=Qt- SKt (4) the period of attachment to the present
(a dot above any variable indicates a household is considered, less investment
derivative with respect to time), where a would be undertaken, and the age of
is the rate by which the stock of human entry into the labor force would be lower.
capital deteriorates. Equations (1) and The problem posed here is suitable
(4) imply that a unit of capital can be for treatment by the techniques of opti-
used from the moment that it is pro- mal control. Three phases are suggested
duced. by the constraints on s, the fraction of
Investment costs have two compo- human capital, or time, allocated to the
nents: production of human capital: (i) The
available stock of human capital Kt,
It = aostKt+ PdDt, (5)
even when fully allocated to produce
that is, (a) opportunity costs, or "fore- human capital, is not large enough to
gone earnings" (the value of the produc- provide the flow of services demanded
tive services withdrawn from the mar- given the relevant prices. The upper
ket), and (b) the direct costs of pur- bound on s is thus an effective con-
chased goods and services. Minimizing straint. (ii) The available stock is large
It with respect to st and Dt, subject to enough to supply the services demanded
 HUMAN CAPITAL AND THE LIFE CYCLE OF EARNINGS 355

and more, so that 0 < s < 1 and the the market so that no earnings are real-
services of human capital are truly a ized. Using conventional tools, we shall
variable factor. (iii) The stock of capital first analyze the later phases in which
is too big so that the optimal policy s < 1 and then shall return to discuss
requiresmore disinvestment than is fea- the first.
sible through deterioration, that is, to Differentiating (7) with respect to Q
produce negative quantities of human we get (9), the marginal cost of produc-
capital. Here st is constrained by its ing human capital; it is a rising function
lower bound.2 of the quantity produced starting from
The first phase is by definitiona period the origin and is independent of the size
of the existing stock of capital of the in-
in which no human capital is allocated to
dividual.
2 In terms of the techniques developed by Pon-
tryagin et al., the maximization of (8) subject to (2), =
_
9 2(_1?#2)
(I3Pd
MCt
(3), and (4) involves the maximization of the Hamil- 00i3 0y32ao/ (9)
tonian (Qt It1/(t6+02) -

H:e-r[(l - s)aoK - PdD]

(1')
+ qQ - K), The value at time t of acquiring an addi-
tional unit of human capital is the dis-
where q is the discounted shadow price of investment
in human capital counted value to that time of the addi-
tions to earnings that the undepreciated
aH
= -e-rt(l- )ao part of this unit will bring about. This is
the "demand price" of human capital,
given by (10). This price is
T
In terms of current prices ,6 - qert
Pt= aof e-(r+5)^dv
i'= -( 1-s )ao
(10)
ao [1-(+)(T-t)
r +
i
- &[I:Q-(a6+ r) ('
independent of the number of units add-
Transversality condition: ed or the existing stock. It is a declining
(3') function of time because of the presence
41(T)K(T) = 0.
of the finite horizon.3 The optimal pro-
First-order conditions for the maximization of (1'): duction of human capital is determined
aH 3 ( 4') by equating the marginal cost to the
=-aoK e-rt + V/e-r Q >0.
C)S ~~~~~S price. Equating (9) and (10) we can solve
-H D for Q as shown in equation (11).
-rt 2
Pd ere+ lIertI2 O. (5') 3 In terms of the framework of n. 2 we are operat-
O9D D
ing in the phase where (4') holds exactly as equality.
A systematic analysis of a variant of this problem, Substituting (4'), (5'), and (2) into (2'), and given
using the techniques of optimal control, is pursued (3'), we get a differential equation to which (10) is a
in an unpublished note by Eytan Sheshinski. solution.

00 02/ 19112
(t =00 )(01+i2)/(l-i,-iO) 0j2ao [ I1-e -(r+b)( T- t) ] (102/ -1112
Q-j -lPde 11)
= NE1 - e (r+S)( T-0)(8i+ft2)/(1-flif2) ?0
356 YORAM BEN-PORATTI

The production of human capital, Q, in ao, the price of the services of human
the gross additions to the stock, are al- capital, do not affect the quantity pro-
ways positive except when t = T. At duced. The obvious reason is that in this
the date of compulsory retirement, T, case, when the services of human capital
the human capital loses its value, P(T) are the only input, a rise in ao raises mar-
= 0 (see eq. [10]). No production is un- ginal cost by exactly the same amount
dertaken at period T, and the stock of that it raises the value of a unit of human
human capital is reduced by sK (see eq. capital. Needless to say, all this discus-
[3]). The point I = T lies in the third sion refers to alternative stationary price
phase. The reason why only this point levels when the current prices are ex-
lies in this phase is that for any earlier pected to prevail throughout the indi-
age, t < T, the demand price of human vidual's lifetime. Qt is larger the larger
capital is positive, and the fact that the the parameters of the production func-
marginal cost curve starts from the origin tion Oo, A1, 02; the greater the length of
insures that some positive quantity will economic life T; and the lower the rate
always be produced. This would be true of deterioration, which has exactly the
of any production function homogeneous same effect as the rate of interest.
of a degree less than 1 in these inputs. As t rises, the flow of Q produced de-
clines, as indicated by (12), the deriva-
MC tive of Q with respect to t:

i-P2
Q = 1-A1-
vP3
- e-(r+5)( T-t) ] [(IIl-)/(1O2)I1 ( 1 2)
.: "P4
X e(r+5)(T[-(r-+t ) 1? 0
What we have here is a marginal cost
Q4 Q3"Q1 curve as a function of Q starting from
Q2
the origin and remaining stationary
FIG. 1
through time, and a perfectly elastic
The market prices that enter into the demand curve which slides down with
determination of the optimal production time, thus creating a pattern of positive
of human capital Q are: r, the rate of but declining optimal quantities of Q to
interest; ao, the rental on human capital; be produced (Fig. 1). If the horizon were
and Pd, the price of purchased inputs. infinite, the demand curve would not
Differentiating (11) with respect to these slide down, and there would be at this
shows that Q varies inversely with the phase one stationary rate of production
rate of interest and with the price of of human capital (given by N in eq. [11}).
purchase inputs and directly with the Note that we are saying something about
rental on human capital. It is, however, the gross additions to stock of human
only the ratio between the last two that capital but not about the net additions,
is important. The elasticity of Q with which depend also on the rate of depre-
respect to the relative price ao/Pa is 32/ ciation and the size of the existing stock.
(1 - /1 - 12)- When 12 = 0, that is, If the initial stock is very high, net addi-
when purchased inputs do not enter into tions may be negative from an early age.
the production of human capital, changes If we think of the normal case as one
 11UMAN CAPITAL AND THE LIFE CYCLE OF EARNINGS 357

where the capital stock does rise over a capacity. The existence of a downturn in
period, eventually as gross additions be- observed earnings is here only a conse-
come very small and the stock becomes quence of depreciation (recall that inter-
large this must be reversed, and toward actions with "leisure" are excluded by
the end of life, T, the stock will decline, assumption). If depreciation is zero,
if there is any deterioration. there is always, except at T, an increase
What are the implications of this pat- in the three types of earnings, and at
tern of investment for the life cycle of each point in time their rank by rate of
earnings? Disposable earnings were de- change will be the reverse of their rank
fined as the difference between earning by level.
capacity and investment outlays; their From the second derivative of ob-
rate of change over time is given by (13). served earnings with respect to time (16),
we learn more about the shape of the
E1t-aoKt-It . (13) life cycle of earnings:
Let us define "observed earnings" E i (+ 3)AICt 01
C)

to be actual earnings realized in the labor =- e-(r+)( T-t)

031+132

market (14).
| 1 - 1 e- (r+b) (T+t)0 16 )
flt::-:aoKt- AIt (14)
/31?102

- ao6Qt + aoa2Kt
They are larger than disposable earnings
by the direct costs PdD and smaller than If there is no deterioration, that is, a = 0,
earning capacity by foregone earnings. the ever rising curve of observed earnings
Their rate of change over time is given is always concave from below. This is al-
by (15). so true of the rising portion of the curve
when a > 0. If /1 + 02 > 2, there will be
?aokt - - It = aoQt a certain range in the vicinity of t = T
(15) where the curve is convex from below.4
- +:-(AICt)Qt-aobKt All these qualitative results hold for any
production function homogeneous of a
I is always negative (except when t = degree smaller than 1 in sK and in D.
T), so that the change in observed earn- 'Q is negative, and the term in braces is positive:
from the equality between (11) and (12)
ings is always algebraically larger than
the change in earning capacity and
ao= Arc.
smaller than the change in disposable I1- e-(r+b)( T-0)Al

earnings The right-hand side of this equality is here multi-

plied by a term always smaller than + 1 and de-
(:1 + 02) ducted from the left side so that the difference is
positive and the first term is negative. If 6 = 0 and
Thus, the curve of observed earnings ex- there is no deterioration, the other two terms vanish,
and the curve of observed earnings is always concave
aggerates the rate of increase of earning from below. This is also true of the curve of dis-
capacity when the latter increases and posable earnings with 1 substituted for 01/(O3 + 32).
understates its decline when it declines. When a > 0, concavity is assured for the range
In particular, when we observe an indi- where K = Q - 6K > 0 If Al + 02 > a, the first
term tends to zero as t -* T; in the vicinity of t = T.
vidual at the peak of his earnings he is K < 0, so in this case the earnings curve will have
already past the peak of his productive there a shape that is convex from below.
358 YORAM BEN-PORATH

The preceding discussion relates to the ules independent of the future levels of
phase where the optimal s, the fraction Q, here only a complete, dynamic pro-
of human capital or time allocated to the gram in which the effects of present ac-
production of human capital, is small- tions on future condition are explicitly
er than 1. As indicated before, the maxi- taken into account will generate an opti-
mization of (8) may, however, require mal path of investment.
that there will be a phase of complete Phase (i) is identified by the absence
specialization in the production of human of earnings. Another distinguishing char-
capital in which s = 1 and the size of acteristic of this phase is that the relaxa-
the stock is an effective constraint. tion of the constraint Kt which lowers
Movement along the marginal cost the marginal cost curve may induce a
curve (7) by increasing Q implies an in- pattern of increasing investment with
crease in the required sK. Once s = 1 is age, both through the allocation of the
reached, larger rates of production can be services of a growing stock of human
achieved only by combining more pur- capital for the purpose and through
chased inputs with a fixed flow of serv- higher direct expenditure on D, justified
ices of human capital, thereby increasing by the higher marginal productivity of D
costs at a higher rate than is implied by that the increased K brings about. A
(7). The marginal cost curve described phase of increasing direct costs in the
by (7) is a long-run envelope from which early period when no earnings are real-
steeper marginal cost curves rise up, ized is certainly not inconsistent with
corresponding to alternative levels of the the real world.5
available stock of human capital. When The age t* when the individual enters
people are young the value of a unit of the second phase and leaves the first is
human capital is only negligibly affected of great importance, because this is the
by the finite horizon T, so demand for date at which positive earnings begin to
human capital is relatively high. On the be realized. From the preceding discus-
other hand, the available stock is still sion it should be clear that K(t*) is a de-
small; therefore the marginal cost curve clining function of t*.6 Ko is the initial
rises up from the long-run curve, (7), at endowment of human capital, which,
relatively small output, so that at a young following the definition of K, is propor-
age production is likely to occur at out- tional to the maximum earnings the in-
puts where s = 1. dividual can realize when he "starts."
Note that the demand price given by All other parameters given, the larger
(10) is now only the lower limit of the this initial endowment, the earlier will
true shadow price of human capital. An be the date t* at which specialization
increase in the stock of human capital stops, noting that Ko is an initial endow-
implies a reduction in future costs of pro- ment only in terms of the capacity to
duction of human capital. The precise earn, while we hold constant the capacity
value of this depends on the future 5 An explicit solution of our problem for the case
quantities of Q that will be produced in which 02 = 0 is simple but not interesting. The
between t and t*, the date of transfer interesting aspects of the problem derive from the
from the first to the second phase. While possibility of using purchased inputs.
6 This can be verified by substituting s = 1 into
in phase (ii), where 0 < s < 1, we were
(4'), letting it hold as an equality, and substituting
able to determine the optimal Qt by the into it also the value of Q and Qkusing (2), (5'), and
intersection of demand and supply sched- the solution of (2') referred to in n. 3.
 HUMAN CAPITAL AND THE LIFE CYCLE OF EARNINGS 359

to increase human capital thus defined, o and 1. We shall first regard s as a

which is reflected in the O's of the pro- measure of the allocation of time to the
duction function. production of human capital, although
this interpretation is not necessary.7
II. THE PRODUCTION FUNCTION
At any time t the stock of capital Kt
ROLE AND IMPLICATIONS is given. If we were to assume that any
The technology which the individual activity the individual engages in pro-
faces when he makes decisions about in- duces either human capital or earnings
vesting in himself is a complicated sys- but not both, then the allocation of time
tem of technical and institutional rela- between these two types of activities is
tionships covering a wide spectrum of also the allocation of the services of the
activities including formal education, ac- existing stock of human capital. The
quisition of skills on the job, child care, larger the stock of human capital, the
nutrition, health, etc. By writing down a larger the earnings per unit of time that
simple production function of the sort the individual could get in the market
used here we are attempting, not to re- and therefore the higher the foregone
produce this system, but only to provide earnings from diverting a unit of time
a framework within which some of the away from the market (see Becker, 1965).
possible characteristics of the technology Whether this should or should not affect
can be considered and their implications the relative attractiveness of non-market
studied. activities hinges on whether the change
that made an hour in the market more
A. PURCHASED INPUTS, TIME, AND THE rewarding also made more productive an
SERVICES OF HUMAN CAPITAL
hour outside the market, in our case an
Both the theory and the measurement hour of producing human capital.
of investment in human capital (see The question is whether the real pro-
Schultz, 1963) emphasize the importance duction relation involved is stable in
of foregone earnings alongside direct terms of time or in terms of some other
cost. A composite surrogate for direct variable. The way we defined human
cost here is D, the index of purchased capital was to make the production of
inputs which stands for anything from earnings stable in terms of its services,
tuition to vitamin pills. The presence of rather than in terms of time as such. By
opportunities for some substitution be- making (sK) the relevant input in the
tween purchased and own inputs has a production function of human capital
"smoothing" effect on behavior and we are also asserting that this other
helps the individual overcome the con- process is stable in terms of the services
straints of his limited time and some- of human capital rather than in terms of
times modify the effects of its increasing time as such. The main implication of
cost. this formulation in terms of the alloca-
The nature of and the role played by
own inputs require some further discus- 7 Note that time enters in two completely differ-
ent ways-I, which moves the individual along his
sion. In the production function so far life cycle and which is being treated as continuous,
considered, own inputs are represented and se, which is a measure of the allocation of time
by the product sK. K is the total stock at any point in t. If we were to treat t as an index of
discrete periods, and if s were understood as referring
of human capital; s is described as a to the allocation of each such period, this dual mean-
fraction that can take any value between ing probably would have posed no problem.
360 YORAM BEN-PORATH

tion of time is that, when K is higher, the If y1 = 7Y2, we get an expression iden-
higher costs of an hour diverted away tical to (2), with yYreplacing 01. If -Yj>
from the market (because of the higher 72, the marginal cost curve is shifting up-
foregone earnings) are exactly matched ward as K increases. This can be viewed
by the greater productivity of time in as a case where time plays an independ-
the production of human capital, so that ent part in the production of human
the marginal cost curve of producing the capital or, more directly, as a case where
latter is independent of the stock of the larger productivity of time in the
human capital in the range where the market indicated by an increase in K is
constraint on s is not effective. not completely matched by a larger pro-
By arguing that in the production of ductivity in the production of human
human capital the services of human capital, so that the cost of the latter in
capital rather than time as such are terms of the former rises. The case where
relevant we gain some analytical sim- 71 = 72 implies, for example, that the
plicity, because beyond the period of more highly educated person is also bet-
complete specialization in the production ter equipped for learning, so that his high-
of human capital the benefits associated er opportunity cost is matched by the
with acquiring a unit of human capital, greater amount of skills that he can ac-
the addition to future disposable earn- quire per hour. If this is not so, then the
situation that we are describing now is
ings, do not depend on the future alloca-
relevant. A limiting case would be one
tion between the market and the produc-
in which human capital does not at all
tion of human capital. This is the reason
affect the ability to produce more human
why the dynamic programing problem capital,
72 = 0.
that was referred to in relation to phase In the basic model considered before,
(i) degenerated into a much simpler de- the decline over time in investment
cision problem in phase (ii). Analytical beyond the period of complete specializa-
convenience is, however, no substitute tion is brought about by the downward
for relevance. In a general production drift, due to the approaching horizon 7'
function both time as such and human of the demand function, along a station-
capital would appear as inputs, and if ary, upward-sloping marginal cost curve.
we consider again the Cobb-Douglas In the situation now described, as K in-
case, s and K may appear with different creases the cost curve shifts upward. In
coefficients. Thus, consider (17): young age, when T - t is large, the
quantitative effect of changes in I on the
Q= o'i' D demand price is small, and the shifts in
= /3oSY172 ( sK)72DO2 ; (17) the cost function may be more important
for changes in investment than in the
0 < Yl Y2,2, < 1 .
downward drift of the demand curve.
T1he corresponding marginal cost curve The role of purchased inputs here is also
is (18): clear-the larger the 02, the smaller the
effect of K on the marginal cost.
MCt = ao (I/ Pd )2/(7+1?2) The other case, in which Y2> -y, also
do71 02 ao cannot be ruled out. Here capital accu-
(18)
+ reduces the cost of producing
x Qt mulation
human capital, and it is possible even in
 HUMAN CAPITAL AND THE LIFE CYCLE OF EARNINGS 361

phase (ii) to have a stretch of time over capital finite. One form in which con-
which investment rises rather than de- straint on the rate of investment could
clines (the downward shift in the costs come about is from individual capacity
being more important than the declines limitations solely. Thus, consider the
in demand). The corresponding life cycle case in which (2) 01 + 12 = 1, j 1 > 0,
may then have an early convex portion. and 12 > 0. In every period the available
Eventually the declining marginal pro- stock of human capital Kt is given. Costs
ductivity of K (due to 72 < 1) and the would be constant up to that level of
growing effect of the approaching horizon output which implies s - 1, that is,
T will turn the rise in investment into
a decline, and the life cycle will have the Q (= Pd 01
familiar S-shape.
For larger outputs marginal costs would
B. THE ROLE OF RISING COSTS rise (if 12 > 0) or become infinite (if
With a homogeneous stock of human 12 = 0). In either case, given the nature
capital, the services of which are sold at of the demand for human capital, the
a fixed price in a competitive market, optimum rate of investment would al-
both the determination of a finite de- ways imply a value of 1 or 0 for s > n,
sired stock and the speed in which the the allocation of all the services of the
available stock is adjusted to the desired available stock of human capital either
level depend on the cost of acquiring to the production of human capital or to
human capital. If human capital could the labor market, but not to both. Be-
be acquired at a fixed (or declining) cost cause of the declining demand over time
without limitation, that is, if the supply for human capital, a period of complete
schedule were perfectly elastic or declin- specialization in the production of human
ing, the desired stock would be either capital (phase [i]), if it comes, must pre-
zero or infinite, and the optimal adjust- cede the period of complete specializa-
ment would be instantaneous. In the case tion in market work (phase [iii]). The
of the demand for tangible capital, costs resulting life cycle of (observed) earnings
of adjustment are sometimes introduced would therefore have a portion of zero
to explain investment as a function of earnings and then a jump to some posi-
the interest rate. In the case of the aggre- tive level, which would be stationary in
gate economy, it is the rising cost of in- the absence of depreciation and declining
vestment goods in terms of consumption in its presence.
goods that provides the negative slope Neither this implied life cycle of earn-
of the aggregate demand function of in- ings nor the behavior to which it is re-
vestment (the marginal efficiency of in- lated is supported by what we know of
vestment curve) and a finite rate of in- the real world. The basic model allows
vestment. for an initial period of complete speciali-
The nature of human capital, the fact zation in the production of human cap-
that it has to be produced by the indi- ital and, correspondingly, a period of
vidual, makes for some similarity in the zero observed earnings. But if we assume
considerations involved and provides a that the sum of the production elastici-
natural basis for dealing with the ques- ties of the variable inputs is smaller than
tion of what makes the rate of invest- 1 (01 + 12 in eq. [2] and yl + 12 in
ment and the attained stock of human eq. [17], the more general formulation),
362 YORAM BEN-PORATH

marginal cost rises continuously even pure work and pure learning; jointness is
before the capacity constraint is reached. too prevalent to be excluded. We can
This is the source of the existence of a still think of the individual as being faced
phase (ii), a period in which 0 < s < 1, with a production frontier indicating the
when the individual engages simultane- possible combination of flows of earnings
ously, or alternately, in work in the mar- and of human capital that he can pro-
ket and in the production of his human duce. Movements along this frontier,
capital. (A more complicated functional however, would not necessarily represent
form could have allowed increasing re- shifts between pure activities but, rather,
turns to scale at small outputs and even- represent movement between activities,
tually decreasing returns. This may be each offering a different mix of earnings
more plausible and still would provide and additions to productive capacity.
the eventual check on the rate of invest- Shifts along this frontier are represented
ment.) by the control st, and they may involve
In this framework, where there is not a change in occupations, of jobs within
"automatic" growth in earnings, in- the occupation, or of function with the
crease in earning can come only from the job. The less numerous and close together
allocation of more services of human are these alternative combinations, the
capital to the market. This can be either smaller is the justification for the con-
because of a reduction in what is allo- tinuous differentiable frontier implied by
cated to investment (st) or because of an our formulation. Discontinuities and
increase in the total stock available (K). kinks in this production frontier are
In order for some positive earnings to be translated into kinks and jumps in the
observed there must be some work in the life cycle of earnings. In either case, it
market; in order for earnings to decline is clear that st becomes an empty con-
there must be some investment that cept once it has lost its link with the ob-
either increases from period to period the servable phenomenon of the allocation
total available stock or, by itself declin- of time.
ing, raises the fraction allocated to the
III. THE MODEL AND AGGREGATE
market. The closer the sum of the pro-
EARNING PROFILES
duction elasticities of the variable factors
to 1, that is, the closer we get to constant The preceding section clarifies the role
returns to scale in terms of the variable played in this framework by increasing
factors, the more concave is the life cycle costs in the explanation of a gradually
of earnings; the case of constant marginal rising portion in the individual life cycle
cost in which the life cycle of earnings is of earnings. The purpose of the present
a one-step function is the limiting case. section is to show that even within this
Throughout the preceding discussions framework an explanation of cohort life
st has been interpreted as a parameter of cycles of earnings, or of the cross-section
the allocation of time. If there are, how- profile of earnings, can be provided with-
ever, activities in which earnings and out the assumption of rising costs. Thus,
human capital can be jointly produced, assume for simplicity that 02 = 8 = 0
then Stloses this meaning. We have many and /3 = 1. By the argument presented
examples of learning on the job in which in the preceding section, the individual
the time spent on the job cannot be allo- life cycle will be a stepwise function.
cated in any meaningful sense between The marginal (and average) cost of
 HUMAN CAPITAL AND THE LIFE CYCLE OF EARNINGS 363

producing human capital is (ao/to), fo the optimal path of accumulation of

differing among people. The higher is Oo, human capital and the life cycle of earn-
the later comes the date t* when the ings. The concepts and parameters in-
downward-drifting demand price crosses troduced raise again some familiar prob-
it from above-which is the date when lems and are used to focus attention on
people jump from observed earnings of some others. In lieu of a summary we
zero to their earning capacity. By equat- shall now refer to a few of these.
ing aco o with P (10) we get (19): 1. The particular definition of human
capital K used is a measure of a quantity
=T+- ln (1--J . (19) of a source of productive services. It is
r \13o
the stock that produces labor services in
The time path of the stock of human "standard units" and is thus the ana-
capital is given by logue of "machines" in the case of tangi-
Kt= Koefot for 0< 1< 1* ble capital. This should be sharply dis-
(20) tinguished from Wt (defined in [8]),
Kt = Koefoe* for t* < t < T. which is the value of the individual as a
productive factor. It is affected by a
Observed earnings are zero for 0 < t < t* broader view of the individual's produc-
and are equal to earning capacity after- tivity, by his durability, and by the rate
ward. of interest, and it can be misleading as a
Et= aoKoedot* quantity measure of an input in any
given point in time. Being a part of the
- aoKoefot T+(1/r) ln[1-(r/fo)]J(21
individual's net worth (and the price that
Using (21) we can see that, given he could get for himself in a competitive
aoK0, earnings on the time of "emer- slave market), it is probably better de-
gence," 1*, and beyond it, are an in- scribed as "human wealth."
creasing convex function of 1*. If in a 2. The speed of adjustment, the rate
given cohort Ko and i3oare not negatively at which the individual increases the
correlated, we can expect to observe a stock of human capital, determines the
rising curve as we follow the earnings ultimate size of this stock. The consider-
ations affecting the speed of adjustment
history of the cohort through time (or
and the path of investment with age
as we examine the cross-section profile of
merit an explicit analysis. The mecha-
earnings of a completely static popula- nism
provided here relies on the intro-
tion, with stationary distributions by F0 duction of
dependence between the mar-
and Ko, in a static economy). The exact ginal cost of producing human capital
form of the curve depends on the distri- and the rate of production (a justification
bution of the cohort by 1o; reasonable for this can come from the "learning
distributions by jOocan generate a curve curve" considerations). Beyond a certain
with the familiar S-shape. point the saving in costs from postponing
investment to the next period compen-
SUMMARY AND SOME OPEN QUESTIONS
sates for the loss of the returns that
To the theory of investment in man would have come from getting the unit of
we have added here a production func- capital a period earlier.
tion of human capital and explored some Rising costs result here from a certain
of the implications of its properties for specification of the technology. The
361 YORAM BEN-PORATII

properties of the production function and 7. The three market prices-the rate
the parameters introduced draw atten- of interest r, the rental on human capital
tion to certain distinctions that have to a0, and the price of purchased inputs Pd
be made when the conglomerate of abili- -affect behavior in the expected way.
ties and external conditions is considered These are the parameters of the models
and the implications of its properties are that public policy can most directly at-
explored. tack. The framework presented here may
3. In discussing the "initial endow- prove convenient for mapping possible
ment" of individuals a different role is effects of public policy and their inter-
played by the initial endowment in terms action with ability factors.
of the ability to earn in the market (K9) 8. Various problems of measurement
and the abilities to produce additions to are implied but not directly discussed.
earning capacity. These consist of the One of these is the relation between the
ability to make efficient use of inputs integral of investment costs, (7), over
purchasable in the market, D (expressed stretches of time and the integral of K,
by Oo and 32), and of one's own time or of Q. The problem of measurement in
(/0, /1, or -y, in [17]). A related issue is this model hinges on the measurability
to what extent the individual, as he in- of st, which is a question about the preva-
creases his human capital, K, also raises lence of activities in which human capital
his efficiency in the production of K (in and earnings are jointly produced. When
eq. [17] this is expressed by the distinc- se is an observable phenomenon of the
tion between 7y, and 72). This question allocation of time, the model becomes
is related to the homogeneity of human operational, but, without it, much is lost.
capital. 9. We have abstracted here from un-
4. The possibility of producing human certainty and from capital rationing two
capital in addition to earnings means important considerations in investment
that the "real output" of the individual in human capital which should be incor-
consists of earnings plus the value of hu- porated in a complete analysis.
man capital produced, when the latter is 10. A natural extension of the two-way
evaluated by its shadow price. This will choice analyzed here between activities
be a measure of by how much total in- producing current earnings and invest-
activity at a given period would affect W. ment in human capital would be to deal
5. Not much has been said of 8, with the three-way choice involving also
the rate of deterioration. Consideration the allocation of time for activities of a
ought to be given to cases where a is consumptive nature, following the ap-
negative rather than positive as assumed proach of Becker (1965).
here and also to the dependence of de- The reader probably does not have to
terioration on the allocation of human be told how simplifying many of the
capital to different activities. assumptions are and how many addi-
6. The horizon T is treated here as tional aspects of investment in human
exogenous. Opening the analysis to in- capital and of the determination of the
clude leisure will, of course, make retire- life cycle of earnings are relevant. The
ment endogenous, but even if T is the main purpose of this paper was to raise
date of death there can be types of in- some more questions rather than provide
vestments that will affect that date. definite answers.
 HUMAN CAPITAL AND THE LIFE CYCLE OF EARNINGS 365

REFERENCES
Becker, G. S. "Investment in Human Capital: Friedman, M., and Kuznets, S. Income from
A Theoretic Analysis," J.P.E., LXX, No. 5, Independent Professions. New York: Na-
Part 2 (October, 1962), 9-49. tional Bureau of Economic Research, 1945.
. Human Capital. New York: Columbia Mincer, J. "Investment in Human Capital
Univ. Press (for the National Bureau of and Personal Income Distribution," J.P.E.,
Economic Research), 1964. LXVI, No. 4 (August, 1958), 281-302.
. "A Theory of the Allocation of Time,"
. "On-the-Job Training: Costs, Returns,
Econ. J., Vol. LXXV (September, 1965).
and Some Implications," ibid., LXX, No. 5,
. "Human Capital and the Personal Dis-
tribution of Income: An Analytical Ap- Part 2 (October, 1962), 50-79.
proach." New York: National Bureau of Schultz, T. W. The Economic Value of Educa-
Economic Research, October, 1966 (mimeo- tion. New York: Columbia Univ. Press, 1963.
graphed).