Download - Investigación en Capital Humano Becker
-
8/17/2019 Investigación en Capital Humano Becker
1/42
Investment in Human Capital: A Theoretical AnalysisAuthor(s): Gary S. BeckerSource: The Journal of Political Economy, Vol. 70, No. 5, Part 2: Investment in Human Beings(Oct., 1962), pp. 9-49Published by: The University of Chicago PressStable URL: http://www.jstor.org/stable/1829103 .
Accessed: 02/02/2011 05:40
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 unlessyou 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
http://www.jstor.org/action/showPublisher?publisherCode=ucpresshttp://www.jstor.org/stable/1829103?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/action/showPublisher?publisherCode=ucpresshttp://www.jstor.org/action/showPublisher?publisherCode=ucpresshttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/1829103?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=ucpress
-
8/17/2019 Investigación en Capital Humano Becker
2/42
INVESTMENT IN HUMAN
CAPITAL:
A
THEORETICAL ANALYSIS'
GARY S. BECKER
Columbia
University nd
National Bureau of Economic
Research
I. INTRODUCTION
SOME
activities
rimarilyffect u-
ture well-being,while others have
their main impact in the present.
Dining is an example of the latter,while
purchase of a car exemplifies he former.
Both earnings and consumption an be
affected:on-the-job training primarily
affects arnings,
new
sail boat primari-
ly affects consumption, and a college
education is said to affect oth. The ef-
fectsmay operate
either
hrough hysical
resources, uch as a sail boat, or through
human resources, uch as a college edu-
cation.
This
paper
is
concerned with
activities that influencefuture real
in-
come
through
he
mbedding
f
resources
in people. This is called investing in
human capital.
The many ways to invest include
schooling, on-the-job training, medical
care,
vitamin
consumption,
nd
acquir-
ing information about
the
economic
system.They
differn
the relative
ffects
on
earnings
and
consumption,
n the
amount
of resources
typically nvested,
in the size of returns, nd
in
the
extent
o
which the connection between invest-
ment
and return
s
perceived.
But all
im-
1
am
greatlyndebted
o the Carnegie
Corpora-
tion ofNew
York for
he
support
givento the Na-
tional
Bureau
of
Economic
Research o study nvest-
ment
n
education
nd other
kinds
ofhuman
capital.
I benefited
reatly rom
many discussionswith my
colleague
JacobMincer,
nd also withotherpartici-
pants
in
the
Labor
Workshop
f
ColumbiaUniver-
sity.Although
many persons
offered aluable com-
ments n thedraft
repared or he conference,
am
especially indebted to the detailed comments of
Theodore
Schultz,
GeorgeStigler, nd Shirley
John-
son.
prove the physical nd mentalabilitiesof
people and thereby raise real income
prospects.
People differ substantially in their
economic well-being,both among coun-
tries and among familieswithin a given
country.For a while economistswere re-
lating these differences rimarily o dif-
ferencesnthe amount of physicalcapital
since richer people had more physical
capital than others. It has become in-
creasingly vident,however, rom tudies
of income
growth2 hat factors
other
than physical resources play a larger
role
than formerly elieved, thus focus-
ing attention on less tangible resources,
like the knowledgepossessed. A concern
with investment in human capital,
therefore,
ies in
closely
with
the new
emphasis on intangible resources and
may
be
useful
n
attempts
to
understand
the
nequality
n
income
among people.
The original aim of my study was to
estimate
the
money rate
of return to
college
and
high-school
ducation
in
the
United States.
In
orderto set these esti-
mates
in
proper
context
I
undertook a
brief ormulationfthetheory f nvest-
ment in
human capital.
It
soon
became
clear to
me, however,
that more
than a
restatement was called for: while
im-
portant
and
pioneering
work
had been
done on
the
economic return
o
various
2
The evidencefor the
United States
appears to
showthat the
growth
n
capital
per capita
explains
onlya
small
part
of the
growthn per capita
income
and that
the growth n
"technology"
explainsmost
of t.On this ee
S. Fabricant,
Economic rogress
nd
EconomicChange:34thAnnualReport f Ie National
Bureau
of EconomicResearch
New
York: National
Bureau
of
Economic
Research, 1954).
9
-
8/17/2019 Investigación en Capital Humano Becker
3/42
1()
GARY S. BECKER
occupations
nd education
classes,3 here
have been few, f
any, attemptsto treat
the process of
investing n people from
a general
viewpointor to
work out
a
broad set of empirical implications. I
began
then to
prepare
a general
analysis
of nvestment
n human capital.
As the work progressed,
t
became
clearer nd clearer
thatmuch more than
a
gap
in
formal conomicanalysis
would
be filled,for the
analysis of human
in-
vestment ffered
unified xplanation
of
a wide range of empirical
phenomena
whichhad either
been given
ad hoc
nter-
pretationsor had baffled nvestigators.
Among
hese are the following:
1)
Earn-
ings typically
ncrease with
age at a de-
creasingrate. Both
the rate of increase
and
the
rate of
retardation
tend to be
positively
elated
to the level of
skill. 2)
Unemploymentrates
tend to be nega-
tively related
to the level of
skill.
(3)
Firms
in
underdeveloped
countries
ap-
pear
to be
more
"paternalistic"
toward
employees han those ndevelopedcoun-
tries. (4) Younger
persons change jobs
morefrequently
nd receive
more school-
ing and on-the-job
trainingthan older
personsdo. (5)
The distribution
f earn-
ings
is positively
skewed, especially
among professional
and
other skilled
workers.
6)
Abler
personsreceive
more
education
and
other
kinds of training
than others. 7) The division of labor is
limitedby the
extent of the
market. 8)
I
In addition to the earlierworks
f
Smith,Mill,
and Marshall, ee
H.
Clark,
Life Earnings
n
Selected
Occupations
n the
U.S.
(New
York:
Harper
&
Bros.,
1937); J.
R.
Walsh, "Capital Concept Applied to
Man," Quarterly ournal of Economics,February,
1935; M. Friedman nd S. Kuznets, ncomefromn-
dependentrofessional ractice New
York:
National
Bureau of EconomicResearch, 1945);
G.
Stigler nd
D.
Blank,The
Demand and
Supply ofScientific er-
sonnel New York: National Bureau of Economic
Research,1957); and T. W. Schultz, Investment n
Man:
An
Economist's View," Social ServiceReview,
June,
1959.
The
typical nvestor
n human
capital
is
more mpetuous
and thus
more ikely
to
err han
s the typical
nvestor
n
tangible
capital.
What
a diverse nd possibly
ven
confusing rray Yet all these as well as
many
other
importantempirical
mpli-
cationscan
be derived
fromverysimple
theoretical
arguments.
The purpose
of
this paper
is
to
set
out
these arguments
in some generality,
with the
emphasis
placed on
empirical implications,
al-
though
little
empirical
material
is
pre-
sented. My
own empiricalwork
will ap-
pear
in a later study.
First, a lengthydiscussionof on-the-
job
training
s
presented
nd
then,
much
more
briefly,
iscussions f nvestment
n
schooling, nformation,
nd health.
On-
the-job
training
s
dealt
with so elabo-
ratelynot because
it is more
important
than
other
kinds
of nvestment
n
human
capital-although
its mportance
s
often
underrated-but
because
it
clearly
llus-
trates
the effectof human
capital
on
earnings,employment, nd other eco-
nomic
variables.
For
example,
the
close
connection
between
foregone nd direct
costs
or
the
effect f
human
capital on
earnings
at
different ges is vividly
brought
ut.
The extended discussion
of
on-the-job
trainingpaves
the
way
for
much briefer iscussions
f
otherkinds
of
investment
n
human
beings.
II. DIFFERENT KINDS OF INVESTMENT
A. ON
THE
JOB
Theories
of firmbehavior,no matter
how
theydiffer
n
other
respects, lmost
invariably gnore
he
effect f
the
produc-
tive process
itself on workerproductiv-
ity.
This
is
not
to say
that no one
recognizesthat productivity
s
affected
by
the
ob
itself;
but
the
recognition
as
not
been
fonnalized, incorporated
nto
economic analysis,
and
its
implications
worked
out.
We
now intend to
do
just
-
8/17/2019 Investigación en Capital Humano Becker
4/42
INVESTMENT IN HUMAN
CAPITAL: A
THEORETICAL
ANALYSIS 11
that,
placing special emphasis
on the
broader
economic mplications.
Many workers
ncrease their produc-
tivity
by earning
new skills
and perfect-
ing old ones while on the job..
For ex-
ample,
the
apprentice
usually learns
a
completely
ew
skill
while
the internde-
velops
skills
acquired
in
medical
school,
and both
are moreproductive fterward.
On-the-job
training, therefore,
is a
process
that
raises
futureproductivity
and differs
rom
chool training
n
that
an
investment
s
made
on the
ob rather
than
in an institution hat specializes
n
teaching.Presumably,future roductiv-
ity can be improved
only
at a cost, for
otherwise
here
would be an
unlimited
demand
for raining.ncluded
in
cost
are
a value placed
on the
time
and
effort
f
trainees,
the
"teaching"
provided by
others,
nd
the
equipment nd materials
used.
These are
costs
in
the sense
that
they
could
have been
used
in
producing
current
utput
if they werenot used
in
raisingfuture utput.The amount spent
and the duration
of the
trainingperiod
dependpartly on
the type of
training-
more
is spent for a longer
time
on
an
intern
han on an
operative-partly
on
productionpossibilities,
and partly
on
the
demand
fordifferent
kills.
Each employee
s
assumed
to be hired
for specified
ime
period in
the imiting
case this
period pproaches
zero), and
for
the moment both labor and product
markets
are
assumed
to be perfectly
competitive.
f
therewere
no on-the-job
training,wage
rates would be given
to
the firm nd wouldbe independent
f
ts
actions.
A
profit-maximizing
irm
would
be
in
equilibrium
whenmarginal roducts
equaled
wages,
that is,
when
marginal
receipts
equaled marginal
expenditures.
In symbols
MP=W,
(1)
where
W equals wages
or
expenditures
and MP
equals the marginalproduct or
receipts. Firms would not
worry too
much
about the relationbetween labor
conditions in the present and future
partly because workers
were
only
hired
forone
period, nd partlybecause wages
and
marginal products
n
future
periods
would
be independent f a firm's urrent
behavior.
t
can therefore
egitimately
e
assumed
that workershave
unique mar-
ginal
products (for given amounts of
other
nputs) and wages
in
each
period,
which
are, respectively,
he
maximum
productivityn all possible uses and the
marketwage rate. A more
complete set
of
equilibrium onditions
would be the set
MPt=
Wt,
(2)
where t refers
to the
tth
period.
The
equilibrium
osition
or
achperiodwould
depend
only
on the
flowsduring
that
period.
These conditions re alteredwhen
ac-
count s taken of on-the-job raining nd
the
connection
hereby reated between
present nd future eceipts nd
expendi-
tures.
Training might
lower current
receipts
and
raise
current xpenditures,
yet firms
ould profitably
rovide
this
trainingf
future eceiptswere sufficient-
ly raised
or future
expenditures
suf-
ficiently
owered.
Expenditures during
each period need not equal wages,
receipts
need not
equal
the
maximum
possible
productivity,
nd
expenditures
and
receiptsduring
ll
periods
would be
interrelated. The set of
equilibrium
conditions summarized
in
equation (2)
would be
replaced by an
equality
be-
tween the
present
alues of
receipts
and
expenditures.
f
Et
and
Rt
represent
x-
penditures
nd
receiptsduring period t,
and i the marketdiscountrate,then the
equilibriumcondition
can
be written s
-
8/17/2019 Investigación en Capital Humano Becker
5/42
12 GARY S. BECKER
Rn
I
t
n
-I
E t
v
(3)
0
(I +i)
t=o
+
where
n
represents
the
number of
periods, and R, and E, depend on all
other receipts and
expenditures.
The
equilibrium
condition of
equation (2)
has been
generalized, for
if
marginal
productequals wages
in
each
period,
the
present value
of the
marginal product
stream would
have
to
equal
the
present
value
of the
wage stream.
Obviously,
however, he
converseneed not
hold.
If trainingweregiven
only during
he
initial period, expenditures during the
initialperiod
would equal wages plus
the
outlay
on
training, xpenditures
during
other
periods
would
equal
wages alone,
and
receipts
during
all
periods
would
equal marginal
products. Equation
(3)
becomes
n-i
MRt
MPO
_
-
O
+1
(
1
+0Z
t
(4)
=WO
k
E
t)tX
where
k
measuresthe
outlay
on
training.
If a new term
s defined,
G=
EM1-Wt
(5)
t=__
(1
+iW
equation
(4)
can be written
s
APo
P+G=Wo+k.
(6)
Since
the term
k
only
measures
he
actual
outlay
on
training
t does
not
entirely
measure
training costs, for
excluded is
the time that
a
person spends on this
training,
ime
that
could
have been
used
to
produce
current
output.
The
differ-
ence between
what could
have beenpro-
duced,
call
this
MPo
and
what is
pro-
duced, MPo, is the opportunity ost of
the
time
spent
n
training.
f
C is
defined
as the
sum of
opportunity
osts and
out-
lays on
training, 6)
becomes
MP'+G=Wo+C.
(7)
The term G, the excess of future
receipts ver future
utlays, s a
measure
of the return
o
the firm rom
providing
training;
and, therefore,
he difference
betweenG and C
measures
the
difference
between he return
rom,
nd the
cost
of,
training.Equation
(7) shows
that mar-
ginal
productwould equal
wages in
the
initial period only
when the return
equals
costs,
or
G
=
C;
it
would
be
greateror less than wages as the return
was smalleror greater
han costs. Those
familiarwith
capital theory
might rgue
that this
generalization
of the
simple
equality between
marginal productand
wages
is
spurious because a full
equi-
librium
would require
equality between
the return
from n
investment-in
this
case,
made on the
ob
and
costs.
If
this
implied that G
=
C,
marginal product
would equal wages in the initialperiod.
There
is
much to be said for
the rele-
vance of a
condition
quating
the
return
from n investment
withcosts, but such
a
condition oes
not
mply
hat G
=
C or
that
marginal
product equals
wages.
The
following
discussion demonstrates
that great care
is
required
n the
applica-
tion of this
condition to
on-the-job
investment.
1. General. Our treatment fon-the-
job
trainingproduced
some
general
re-
sults
summarized
n
equations (3)
and
(7)
of
wide
applicability,
ut more con-
crete
results
require more
specific
as-
sumptions. n
this and the
following ec-
tion
two
types
of
on-the-job raining
re
discussed
in
turn:
general
and
specific.
General
training
s useful
n
many
firms
in
addition to the firm
roviding t, as a
machinist trainedin the
army findshis
skills of
value
in
steel and
aircraft
irms,
-
8/17/2019 Investigación en Capital Humano Becker
6/42
INVESTMENT
IN
HUMAN CAPITAL:
A
THEORETICAL
ANALYSIS
13
or
a doctor trained (interned)
at one
hospital
findshis
skills useful
at other
hospitals.
Most on-the-job
raining
pre-
sumably
increases
the future
marginal
productofworkersn the firm roviding
it,
but general training would also in-
crease theirmarginal
product
in
many
other irms s well.
Since n a competitive
labor
market
hewage
rates paid by
any
firm re determined
y
marginalproduc-
tivities n other
firms, uture
wage rates
as well as marginal
products
would in-
crease to
firms
rovidinggeneral
train-
ing. These
firms ould
capture
some of
the return fromtraining only if their
marginal product
rose by
more than
their
wages.
"Perfectly
eneral"
training
would
be equally
useful in
many
firms
and marginal
products
would
rise
by
the
same
extent
n
all of them.
Consequent-
ly,
wage rates would
riseby
exactlythe
same
amount
as the marginal
product
and
the firmsproviding
such
training
could
not
capture
any of
the
return.
Why,then,do rationalfirms n com-
petitive
labor
marketsprovide
general
training,for why
provide
training
that
brings no
return?
The answer
is
that
firms would provide
general
training
only
if
they
did not have
to
pay any of
the costs. Persons receiving general
training
would
be willing to
pay
these
costs since
trainingraises their
future
wages.
Hence
the
cost as well
as
the
re-
turn from general trainingwould be
borne
by
trainees,not
by firms.
These and other
mplications
of gen-
eral
training
can
be
more
formally
demonstratedwith
equation
(7). Since
wages
and
marginal
products
are raised
by
the same
amount,
MP,
must
equal
W,
for
all
t
=
1, .. .n-1
,and
there-
fore
__
__
0MP- t=?.
( 8)
(1+it
-
t
t=1
(
8
Equation
(7) is reduced
to
MP,=W
WO+C,
(9)
or
VO=MP,
-C.
(10)
In terms
of
actual
marginal
product
MPo=Wo+
k,
(9')
or
Wo=MPo-k.
(10')
The wage
of
trainees
would
not equal
theiropportunity
marginalproduct
but
would
be less by
the
total
cost of
train-
ing. n otherwords, mployeeswouldpay
for general
training
by
receiving
wages
below
their
current
opportunity)
pro-
ductivity.
quation
(10)
has
manyother
implications,
nd the
rest
of this
section
is
devoted
to developing
the more
im-
portant
ones.
Some might
rgue
that
a
really
"net"
definition
f marginal
product
obtained
by
subtracting
training
costs
from
"gross" marginal product must equal
wages
even for
trainees.
Such
an
inter-
pretation
of
net productivity
ould
for-
mally
ave the
equality
between
marginal
product
nd wages
here,
but later
show
(pp.
18-25)
that it
cannot always
be
saved.
Moreover,
regardless
f which
n-
terpretation
s used,
training
ostswould
have
to be
included
n
any
study
ofthe
relation
between
wages
and
productivity.
Employeespay forgeneralon-the-job
training
by
receiving
wages
below what
could
be received
elsewhere.
Earnings"
during
the
trainingperiod
would
be
the
difference
etween an
income
or
flow
term,
potential
marginal
product,
nd
a
capital
or
stock
term,
training
costs,
so
that
the
capital
and
income
accounts
would
be
closely
intermixed,
with
changes
in
either affecting
wages.
In
otherwords,earningsof personsreceiv-
ing
on-the-job
training
would
be
net
of
-
8/17/2019 Investigación en Capital Humano Becker
7/42
14
GARY S. BECKER
investment osts
and would
correspond
to
the definition f net earnings
used
throughout
his paper, which
subtracts
all investment osts
from"gross"
earn-
ings.Therefore,
ur departure
with this
definition f
earningsfrom he
account-
ing conventions
used for transactions
n
material
goods-which
separate income
from
capital accounts
to prevent
a
transaction
in
capital from
ipso factor4
affectinghe income
side
is not capri-
cious
but is grounded
n
a
fundamental
differenceetween
hewayinvestment
n
material nd human
capital are
"written
off."The underlying ause of thisdiffer-
ence
undoubtedly
s the
widespread
re-
luctance to
treat people as
capital and
the accompanying
endencyto treat
all
wage receipts
s earnings.
Intermixing
he capital
and income
accounts
could
make the
reported
"in-
comes"
of
trainees nusually ow
and
per-
haps
negative, even
though
theirlong-
run or
lifetime
ncomes
were
well
above
average. Since a considerablefraction f
youngpersons
receive
ome
training,
nd
since
trainees would
tend to have
lower
current
nd
highersubsequent
earnings
than other youth,
the correlation
be-
tween
current
onsumption
nd
current
earnings
of
young
people'
would
not
only
be much
weaker
than the
correla-
tion with long-run
earnings,
but
the
4
Of course, Eshift etween ssetshavingdiffer-
ent productivities
ould affect
he income
account
on material
goods even
with current
accounting
practices.
I
I say "young people"
rather han "young
fam-
ilies" because
as J.Mincer
has shown
in
a paperto
be published
n a
NationalBureau
ofEconomic
Re-
search
conference
olumeon
labor
economics),
the
labor-force
articipation
f wives is
positively
or-
related
with thedifference
etween
husbands'
ong-
run and current
ncome.
Participation
of wives,
therefore,
akes
the correlation
etween
family's
current nd a husband's long-runncomegreater
than thatbetween
husband's
currentnd
long-run
income.
signs
of these
correlations
might even
differ.6
Doubt
has been
cast on the
frequent
assertion
that no
allowance
is
made in
the incomeaccounts
for
depreciation
n
human capital.7
A depreciation-type
item s
deducted,
at
least from
he earn-
ings due
to on-the-job
training,
for
the
costwould
be deducted
during he
train-
ing period.
Depreciation
on
tangible
capital
does not bulk
so large
in any one
period
because
it is
usually
"written ff"
or depreciated
duringa period
of
time
designed
to approximate
its
economic
life. Hence human and tangible capital
appear
to
differmore
n the time
pattern
of depreciation
than
in its
existence,8
and
the
effect n
wage
income
of
a
rapid
"write-off"
f
human
capital
is
what
should
often
be
emphasized
and studied.
Our
point
can
be
put
differently
nd
more rigorously.
he
ideal
depreciation
on
a
capital
asset
during
any period
would
equal
its
change
in value
during
theperiod.In particular, f value rose,a
negative
depreciation
term would
have
6
A difference
n signs
s impossible
n Friedman's
analysis
of
consumer
behavior
because
he assumes
that
transitory
nd long-run
that
is,
permanent)
incomes
are uncorrelated
see
his A Theory
f
the
Consumption
unction Princeton,
N.J.:
Princeton
University
ress,
1959]);
we
are
suggesting
hat
they
maybe
negatively
orrelated
oryoung
persons.
I
See,for
xample,
A.
Marshall,
Principles
f
Eco-
nomics8thed.;
New York: Macmillan
Co., 1949);
C.
Christ, Patinkinon Money, nterest, nd Prices,"
Journal
of
Political
Economy,
August,
1957, p.
352;
and
W.
Hamburger,
The
Relation
of Consumption
to
Wealth
and
the Wage
Rate,"
Econometrica,
anu-
ary,
1955.
8
In a recent
paper,
R. Goode
has argued
(see
"Educational
Expenditures
nd
the ncome
Tax,"
in
Selma J.
Mushkin ed.],
Economics
fHigher
Educa-
tion [Washington:
United
States
Department
of
Health,
Education,
and Welfare
(forthcoming)])
that
educated persons
hould
be permitted
o
sub-
tract
from ncome
a depreciation
llowance
on tui-
tion payments.
uch
an allowance
s
apparently
not
requiredfor on-the-job raining osts; indeed,one
might
rgue,
on
thecontrary,
hat
too much
or
too
rapiddepreciation
s
permitted
n such
investment.
-
8/17/2019 Investigación en Capital Humano Becker
8/42
INVESTMENT
IN
HUMAN
CAPITAL:
A THEORETICAL ANAILYSIS
15
to be subtracted or a positive apprecia-
tion termadded to the income from
he
asset. Since training osts would
be
de-
ducted from arningsduring
he
training
period, he economic value" of a trainee
would at first ncrease rather than de-
crease with
age,
and
only
later would
it
begin
to decrease.9
Training has an important effect n
the
relation between earnings
and
age.
Suppose
that untrained
persons
received
the
same earnings regardless
of
age,
as
shown by the horizontal
line
UU
in
Figure 1. Trained persons would
receive
lowerearnings uring hetraining eriod
because training s paid for then, and
higher arnings t later ages because the
return s collected then. The
combined
effect
f paying for nd
collecting
he re-
turnfrom raining
n
thisway
would
be
to make the
age earnings urve
of trained
persons,
shown
by
TT
in
Figure 1,
steeperthan that of untrained
persons,
the
difference eing greater the greater
the cost of,and return rom, he invest-
ment.
Not
only
does
training
make the
curve steeper but, as indicated by
Figure 1, also more concave;
that is, the
rate of increase
in
earnings s
affected
more at younger han at older
ages. Sup-
pose,
to take
an extreme
ase,
that train-
ing
raised
the level of marginalproduc-
tivity
but had
no
effect n the
slope,
so
that themarginalproductivity ftrained
persons
was
also
independent
of
age.
If
earningsequaled marginal
product,
TT
would
merely
be
parallel
to
and higher
than
UU, showing
neither
lope
nor con-
cavity. Since, however,
earnings of
trained
persons
would
be
below
mar-
ginal productivityduring
the
training
9
n
my study
for
the National Bureau
of
Eco-
nomic Research I try to measure the relation be-
tween depreciation nd age forseveral education
classes.
period
and
equal
afterwards, hey
would
rise sharply
at the end of
the training
period
nd
then evel off as shownby
the
dashed line
T'T'
in
Fig. 1), imparting
concave appearance to the curve as a
whole.
In this
extreme
case
an
extreme
concavityappears;
in less extreme ases
the principlewould
be
the
same and the
concavity
morecontinuous.
Foregone earnings
are an important,
althoughneglected,
ost
of much
human
capital
and should
be
treatedon the same
footing
s
direct utlays. ndeed,
all costs
appear as foregone
arnings
to
workers
T
zc/ I,_
U
AGE
FIG. 1
receiving n-the-job
raining; hat is,
all
costs
appear
as
lower
earnings han could
be received elsewhere, although
direct
outlays,
C, may really be an
important
part
of costs. The arbitrariness f the
division
between
foregone and direct
costs and the resulting advantage
of
treating otal costs
as a whole" can
be
10
The equivalence between foregone nd
direct
costs pplies to consumption s well as to nvestment
decisions.
A
household an
be
assumed
o maximize
utility unction
U(K1, X
2,
*
X)
X
i
. X being consumption oods, subject to the
constraint
P-ix=wV
h - E lX1)?
+
Y
where
pi
is the marketprice
of
the th
good,
I14
he
average wage rate, y non-wage ncome,
hI
he
total
-
8/17/2019 Investigación en Capital Humano Becker
9/42
16
GARY
S.
BECKER
further demonstrated by
contrasting
school and on-the-jobtraining.
Usually
only
the direct
cost of
school training s
emphasized, even though the foregone
cost is sometimes as withcollege educa-
tion) an importantpart of the total. A
shiftof
training
from chools
to
on the
job would,however, everse he emphasis
and make all costs appear as foregone
earnings, ven when direct
outlays were
important.
Income maximizing
firms n competi-
tive labor markets would not pay the
cost of general training nd
would pay
trained persons the market wage. If,
however, raining osts were
paid, many
persons would seek training,
few would
quit during he training eriod,
nd labor
costs would be relativelyhigh.
Firms
that did not pay trained persons the
marketwage would have
difficultyatis-
fying heir kill requirements
nd would
also tend
to
be
less
profitable
han other
firms. irms that both paid fortraining
and lessthanthemarketwage for rained
persons
would
have
the worst of
both
worlds,
for
hey
would
attract
too
many
trainees nd too
few trainedpersons.
These
principles
have been
clearly
demonstratedduringthe last few years
in
discussions of problems
n recruiting
military personnel.
The military offers
number fhours vailable for ither
onsumption
r
work, nd hj the numberof hoursrequired o con-
sume a unit ofthe
th good. By
transposing
erms
the
constraint an
be written s
2(p?+Whti))X
=Wh+y
.
The total
cost or priceofconsuming unitofthe th
good s
the
umof
two
components:
he
market
rice
or direct
outlayper unit,
pi,
and
the foregone arn-
ingsper
unit,
Whi.
expect o show nanother aper
that this formulation f
household decisions gives
extremely seful
nsights
nto number f
mportant
economic
problems,
uch as the
choice between a-
bor and "leisure," the effect f price control on
prices, he
roleofqueues, and the
cause ofdifferences
among
ncome lasses
n
priceelasticities
f
demand.
training
n
a wide
varietyof skills and
many-such
as piloting nd
machinere-
pair-are
veryuseful n the
civilian sec-
tor. Training
s
providedduring
part or
all ofthefirst nlistment eriodand used
duringthe remainder
f the first eriod
and hopefully uring ubsequent
periods.
This hope, however,
s thwartedby the
fact that
re-enlistmentates
tend to be
inversely related
to the
amount of
civilian-type skills
provided
by the
military.11ersonswith these
skills eave
the militarymore
readilybecause they
can
receive much higher
wages in the
civilian sector. Net militarywages for
those receiving raining re
higherrela-
tive
to civilianwages
during hefirst han
during
ubsequentenlistment
eriodsbe-
cause training
costs
are
largely paid by
themilitary.Not
surprisingly,herefore,
first-term
nlistments or killed
obs
are
obtained
much more
easily
than
are re-
enlistments.
The
military
s
a conspicuous
example
of an organizationthat both pays at
least part of training
osts
and does not
pay
market
wages
to skilled
personnel.
t
has had,
in
consequence,relatively
asy
access to
"students"
and heavy
losses of
"graduates."
Indeed, its graduates
make
up
the
predominate art
of the supply
n
several civilian occupations.
For ex-
ample,
well
over
90
per
cent of
United
States
commercial
irline
pilots
received
much of their training in the armed
forces.
The
military,
f
course,
s
not
a
commercial organization
judged by
profits
nd losses and
has
had
no
diffi-
culty
surviving
nd even thriving.
What
about
the
old
argument
that
" See
Manpower
Management
nd Compensation
(Washington:
Government
Printing
Office, 957),
Vol. I, Chart
3,
and the accompanying
discussion.
The military ot
onlywants
toeliminate
he nverse
relation utapparentlywould iketo createa strong
positiverelation
because
they
have such a
large n-
vestment
n heavily
trainedpersonnel
see ibid.).
-
8/17/2019 Investigación en Capital Humano Becker
10/42
INVESTMENT IN HUMAN
CAPITAL: A
THEORETICAL
ANALYSIS
17
firms n competitive abor marketshave
no incentive o
provide on-the-job rain-
ing becausetrainedworkerswould be bid
away by otherfirms?Firms that train
workers re supposed to impart external
economies to other firms because the
latter can
use
theseworkersfree of any
training charge. An analogy with re-
search and
development s often drawn
since a
firm
developing a process
that
cannot be patentedor
kept secret would
impart external economies to competi-
tors."2 his
argument nd analogy would
apply
if firmswere
to pay training osts,
for they would suffer "capital loss"
whenever
rained
workerswere bid
away
by
otherfirms.
irms can, however,
hift
training oststo
trainees nd have an
in-
centive
to do so
when faced
with
compe-
titionfortheir
ervices.
The difference
etween
nvestment
n
training
and
in
research and
develop-
ment can be
put very simply.Without
patents
or
secrecy,
firms
n
competitive
industries cannot establish property
rights n
innovations, nd these innova-
tions
become fair
game
for all
comers.
Patent systems
try to
establish these
rights o
that
incentives an
be
provided
to
invest
n
research.
Propertyrights
n
skills,
on the
other
hand,
are automati-
cally vested, for
a skill cannot
be used
without
permission
of the
person pos-
sessing
t.
This
propertyright
n
skills
is thesource ofthe incentive o invest n
training
and
explains why
an
analogy
with
unowneq
nnovations
s
misleading.
2.
Specific.-Completelygeneral rain-
ing increases the
marginal productivity
of
trainees
by
exactly
the
same
amount
in
firms roviding he training s
in
other
firms.
learly some kinds
of training
n-
crease
productivity by a different
12
These arguments an be found nMarshall, op.
cit., pp.
565-66, although he
compares
training o
land-tenure ystems.
amount
in
firms
providing
the
training
than in other firms.Training that in-
creases
productivitymore
in
firms
pro-
viding
it
will
be
called
specific raining.
Completely specifictrainingcan be de-
fined s training hat has no effect n the
productivityof trainees that would be
useful n in otherfirms.Much
on-the-job
training
is
neither completely specific
not
completely general but increases
productivitymore
in
firmsproviding t
and
falls within
he
definition f specific
training.
The rest ncreases
productivity
by at
least as much
in
other firms nd
falls within definition f general train-
ing. The previous section discussed gen-
eral training and this one will cover
specific training.
A
few illustrations
of
the
scope
of
specific training are pre-
sented
before a formal
analysis
is de-
veloped.
The
military offerssome forms of
training
hat are
extremely seful
n
the
civilian sector,as already noted. Train-
ing is also offered hat is only of minor
use
to civilians:
stronauts, ighter ilots,
and missile
men
all illustrate this to a
greater
or
lesser extent. Such training
falls
within
he
scope
of
specific raining
because
productivity s raised
in
the
military
ut not
(much)
elsewhere.
Resources
are usually spent by firms
in
familiarizing ew employeeswith heir
organization,
and
the
knowledge so
acquired is a formof specific training
because
productivity
s
raised more
in
the
firms
cquiring
the
knowledge than
in
other
firms.
Other kinds of hiring
costs,
such as
employment gency fees,
the
expenses
ncurred
y
new
employees
in
finding obs (what Stigler calls in his
paper
in
this
Supplement
the
"costs of
13
To judge
by a sample of firms
ecently na-
lyzed,formal rientation
ourses re quite
common,
at least in large firms see H. F. Clark and H. S.
Sloan,
Classrooms n the
Factories
New
York:
New
York
University ress,
19551,
hap. iv).
-
8/17/2019 Investigación en Capital Humano Becker
11/42
18 GARY S.
BECKER
search"), or the time employed n inter-
viewing, esting, hecking eferences,
nd
in bookkeepingdo not so obviously raise
the knowledge of new employees,
but
they oo are a form f specificnvestment
in human capital, althoughnot training.
They are an investment
ecause outlays
over a short period create distributed
effects n productivity;
hey are specific
because productivity s raised primarily
in the firmsmaking he
outlays; they are
in human capital because they ose their
value whenever mployees eave.
In the
restof this section usuallyrefer nly to
on-the-job pecific raining ven though
the analysis applies to
all on-the-job
specific nvestment.
Even after hiring costs are incurred,
firms suallyknowonly limited mount
about
the ability
and potential of
new
employees. They try to
increase their
knowledge in various
ways-testing,
rotation among departments, rial and
error, etc.-for greater
knowledge per-
mits a more efficient tilizationofman-
power.Expenditures
n
acquiring
knowl-
edge of employee talents
would be a
specific investment if the knowledge
could be keptfrom ther firms, or
then
productivitywould be
raised more
n
the
firms
makingtheexpenditures
han else-
where.
The
effect
f
nvestment
n
employees
on
theirproductivity
lsewhere
depends
on market conditions s well as on the
nature of the investment. Very strong
monopsonists
might
be completely
nsu-
lated from
competition
by
other
firms,
and
practically
all
investments
n
their
labor
force
would
be
specific.
On
the
other
hand,
firms
n
extremely ompeti-
tive
labor
marketswould face a
constant
threat
of
raiding
and would have
fewer
specific
nvestments
vailable.
These
examples convey
some of
the
surprisingly arge
variety
of situations
that come
under
the rubric
f specific
n-
vestment.
This
set is now
treated ab-
stractly
n order that
a
general formal
analysis can
be
developed.
Empirical
situations are brought in again after
severalmajor
implications
f the formal
analysis
have been
developed.
If all
training
were completely
pecific,
the wage
that
an employee
could get
elsewhere
would
be independent
of
the
amount
of
training
he
had
received. One
might plausibly
argue,
then, that
the
wage paid
by
firmswould also
be
inde-
pendent
of
training.
f so, firms
would
have to pay trainingcosts, forno ra-
tional
employee
would
pay
for training
that did
not benefit
him.
Firms
would
collect the
returnfrom
uch training
n
the form f
largerprofits
esulting
rom
higher
productivity,
nd
training
would
be
provided
whenever
the return-dis-
counted
at an
appropriate
rate-was
at
least as
large as
the cost. Long-run
competitive
quilibrium
equires
hat
the
present value of the return exactly
equals costs.
These propositions
an
be
stated more
formally
with
the equations
developed
earlier. According
to
equations
(5)
and
(7)
the
equilibrium
of a
firm
providing
training
n
competitive
markets can
be
written
s
0~
=
E
+0-t
( 11
=
lvo+C
where
C
is
the cost
of
training iven
only
in the initial
period,
MPo
is the
oppor-
tunity
marginal
product
of
trainees,
Wo
is
the
wage
paid
to trainees,
and
Wt
and
MPt
are
the
wage
and
marginal
product
in
period
t.
If
the
analysis
of
completely
pecific
raininggiven
in
the
preceding
paragraph was correct, W
would
always equal
the
wage
that could
-
8/17/2019 Investigación en Capital Humano Becker
12/42
INVESTMENT IN
HUMAN CAPITAL: A THEORETICAL
ANALYSIS
19
be
received
lsewhere,
MPt
-
Wt
would
be the fullreturn n t
from
raining iven
in 0, and G
would be the
present
value
of these returns.
ince
MPo
measuresthe
marginal product elsewhere and WO
would measure the
wage
elsewhere
of
trainees,MP'
=
Wo.
As a
consequence
G
=
C, or,
n
full
equilibrium,
he return
from
raining quals
costs.
Before
claimingthat
the
usual equal-
itybetween
marginal
product nd wages
holds
when
completely pecific
raining
is
considered, he reader
should bear in
mind
two points. The
first s
that
the
equality between wages and marginal
product n the
initial
period nvolves op-
portunity,
ot
actual
marginal product.
Wages
would
be
greater
than actual
marginal product if
some productivity
was
foregone
s
part
of the
training ro-
gram. The
second is that, even if
wages
equaled marginal
product
nitially, hey
would
be
less
in
the
future
because
the
differences between
future
marginal
products ndwages constitute hereturn
to
training
nd are
collected
by
the firm.
All of
thisfollows rom
heassumption
that firms
pay all costs
and collect all
returns.But could not one
equally
well
arguethat
workers ay all
specific rain-
ing
costs by receiving
appropriately
lower
wages initially
and
collect
all
re-
turns
by
receivingwages equal
to mar-
ginal
product ater?
n
terms f
equation
(11), Wtwouldequal MPt, Gwouldequal
zero, and
Wo=MP'-C, just as with
general training.
Is it
more
plausible
that firms
ather than workers
pay
for
and collect
and return rom
raining?
An
answercan be
foundby
reasoning
along
the
following
ines. If a firm
had
paid
for
he
specific
raining
f
a worker
who
quit
to take
another
ob,
its
capital
expenditurewould be
partly
wasted,
for
no furtherreturn could
be collected.
Likewise,
a worker
fired after he had
paid for specific
training would
be
un-
able
to collect any further eturn and
would also suffer
capital loss.
The
willingness
f
workers r firms o pay
for
specific raining hould,therefore,losely
depend on the likelihood of labor turn-
over.
To bring nturnover t this pointmay
seem ike
a deus exmachine ince t is al-
most always
ignored in traditional
theory.
n
the usual
analysis of competi-
tive firms,wages
equal marginal prod-
uct, and since wages and marginal
product are assumed
to
be the
same
in
many firms,no one suffers rom turn-
over.
It
would not matter whether a
firm's abor force
always contained the
same persons or a
rapidly changing
group.Any person eaving onefirm ould
do
equally well
in
otherfirms, nd his
employer ould replace
him
without ny
change
in
profits.
n
other
words,
turn-
over
is
ignored
n
traditional
theory
be-
cause
it
plays no
importantrole
within
theframework f the theory.
Turnover
becomes important
when
costs
are
imposed
on workers or
firms,
which
are
precisely
he
effects f
specific
training. Suppose a
firm
paid all the
specific training
costs of a worker who
quitafter ompletingt.
According
o
our
earlier analysis
he
would have been
re-
ceiving
the
market
wage
and a
new em-
ployee
could be
hired at the same
wage.
If the new employee were not given
training,
his
marginalproduct would be
less
than that
of
the one who
quit
since
presumably training raised the latter's
productivity.
Training
could raise
the
new
employee's
productivity
ut
would
require additional
expendituresby
the
firm.
n other
words,
firm
s
hurt
by
the
departure
of
a trained
employee
because
an
equally profitable
new
employee
could
not
be obtained.
In
the
same
way
an
employee
who
pays
for
pecific
rain-
-
8/17/2019 Investigación en Capital Humano Becker
13/42
20
GARY S. BECKER
ingwould
suffer
loss from eing aid
off
because
he
could
not
find an equally
good
job
elsewhere.
To bring turnover
into
the analysis
of specific
training
s
not, therefore, deus ex machinebut is
made necessary
by
the
important
ink
between
them.
Firms
paying
for specific
training
might take account
of
turnover
merely
by obtaining
a sufficiently
arge return
from those
remaining
to
counterbal-
ance the
oss
from
hose
eaving. The
re-
turn on "successes"-those
remaining-
would,
of
course,overestimate
he
aver-
age return n all training xpenditures.)
Firms could
do even better,however,
by
recognizing
hat
the likelihood
of a quit
is not fixed
but depends
on wages.
In-
stead of merely
recouping
on
successes
what is lost
on failures, hey might
re-
duce the
likelihood of
failure itself
by
offering
igher
wages
after raining
han
could be received
elsewhere.
n
effect,
they would
offer
mployees
ome
of
the
returnfromtraining.Matters would be
improved
n
some respects
but worsened
in others,
for
the higher
wage would
make thesupply
of trainees
greater
han
the demand,and
rationing
would
be re-
quired.
The final
step
would be to shift
some training
osts
as well as returns
o
employees,
hereby
ringingupply
more
in
line with
demand.
Whenthe
final tep
is completed
firms
no
longer pay
all
training ostsnor do theycollect all the
return
but
they
share both with em-
ployees."4
he shares
of each
depend
on
the
relation
between
quit
rates and
wages,
layoff
rates
and
profits,
nd
on
other
factors
not discussed
here,
such as
the
cost
of
funds,
ttitudes
toward
risk,
and desires
for
iquidity.'5
If training
were
not completely
spe-
cific, roductivity
ould ncrease
n
other
firms s well,
and
the
wage
that
could
be
received
elsewhere
would
also
in-
crease. Such
training
an be looked
upon
as
thesum of two components,
ne
com-
pletely
general,
he
other ompletely
pe-
cific, with
the former
being
relatively
larger the greater the effecton wages
in other
firms elative
to the
firms
ro-
viding
the
training.
Since
firms
do
not
pay
any
of
completely eneral
costs and
onlypart
of
completely
pecific osts,
the
fraction
f costs paid
by
firmswould
be
negatively
related
to the importance
of
the general
component,
r positively
re-
lated
to
the
specificity
f the
training.
Our
conclusions
an
be stated formal-
ly in terms of the equations developed
earlier.
f
G
is the present
value of
the re-
turnfrom raining
ollectedby
firms,
he
fundamental
quation
is
MP'
+
G
=
W
+
C.
(12)
If G'
measures
the
return collected by
employees,
the total return,
G",
would
be
the sum of G
and G'.
In
full
equi-
librium
he
total
returnwould
equal
total
costs, or G" = C. Let a representthe
fraction
f the total
return ollected by
firms. ince
G
=
aG"
and
G"
=
C,
equa-
tion (12)
can
be written
s
14
Marshall
was
clearly
aware of
specific
alents
and their
ffect n
wages
and
productivity:
Thus
the
head
clerkin
a business
has an
acquaintance
withmen
and things,
he use
of which
he could
in
somecases
sell at a
highprice
to rivalfirms.
ut
in
other
ases
it
sof
kind
to
be
ofnovalue
save to
the
business nwhichhe already s; and thenhisdepar-
turewould
perhaps
njure
t by
everal
imes
hevalue
of his
salary,
whileprobably
he
could
not
get
half
thatsalary
elsewhere"
op. cit.,
p. 626). (My
italics.)
However,
he overstressed
he element
of indeter-
minacy n
these
wages ("their
earnings
re deter-
mined ..
bya bargain
between
hem nd
their m-
ployers,
the terms
of which are theoretically
r-
bitrary" ibid.,
fn.])
because
he
ignored
he effect
f
wages
on turnover.
15
The
rate used
to discount
osts
and
returns s
the
sum
of a (positive)
rate measuring
he cost of
funds,
(positive
or
negative)
risk
premium,
nd a
liquiditypremium hat s presumably ositive ince
capital
invested
n
specific
raining
s
very
lliquid
(see
the discussion
n
Sec.
IV,
C).
-
8/17/2019 Investigación en Capital Humano Becker
14/42
INVESTMENT
IN HUMAN
CAPITAL:
A
THEORETICAL
ANALYSIS
21
MP'+ aC
=
W
+
C, (13)
or
I == P'- (1 a)C.16 (14)
Employees pay the same
fraction of
costs, 1
-
a, as they collect in returns,
which generalizes
the results obtained
earlier.
For
if
training
were completely
general, a
=
o, and equation
(14) re-
duces to equation (10); iffirms ollected
all the return
from
raining,
=
1,
and
(14) reduces to MP'
=
Wo;
if
0
< a <
1, none of
the
earlier quations
are satis-
factory.
A few
major implications
of this
analysis of specific training are now
developed.
Rational
firms
pay generally
trained
employees
he
same wage
and specifically
trained
employees
a
higher
wage
than
they could get
elsewhere.A
readermight
easily
believe the
contrary,
amely,that
general rainingwould
command higher
wage
relative o alternatives han
specific
training oes, since, fter ll, competition
for
persons
with the
latter
is
apt
to be
weaker than for those
with
the former.
This view, however, overlooks
the
fact
that
general training
raises the
wages
that
could
be received elsewhere while
(completely) specifictraining
does
not,
so
a
comparison
with alternative
wages
gives a misleading impression
of the
absolute ffect
n
wages
of
different
ypes
of training.Moreover,firms re not too
concerned about
the
turnover of
em-
ployeeswith general training nd have
no incentive to
offer hem a premium
above
wages
elsewhere
because
the cost
16
IfG" didnot equal C, these
quationswould
be
slightlymore complicated.
Suppose, for example,
G"
=
G
+
G'
=
C
+
-n
n
>
0 so that the
present
value
of thetotal returnwouldbe greater han
total
costs. Then
G
=
aG"
=
aC
+
an, and
MIP'+ aC+ an = W+C,
or
W
=
AP'- [ 1-a
)C-anl].
of such
training
s borne
entirely y
em-
ployees. Firms are concerned
about the
turnover
of
employees
with
specific
training,
nd a
premium
s offered
o re-
duce their turnover because firmspay
part oftheir
training
osts.
The
part
of
specific
rainingpaid by
employees
has effects imilar
to those
discussed
earlier for
generaltraining:
t
is also paid by a reduction
n
wages
dur-
ing the training
period,
tends
to make
age-earnings profiles
steeper
and more
concave,
etc.
The
part
paid by
firms
as
none
of
these
implications,
ince
current
or futurewageswould not be affected.
Specific,
nlike
general, raining
would
produce certain
"external"
effects,
or
quits
would
prevent irms
rom
apturing
the
fullreturn
n
costs
paid by
them, nd
layoffs
would do the same to
employees.
Note,
however,
that
these are
external
diseconomies
mposed
on
the
employees
or
employers
f
firms roviding
he train-
ing,
not
external
economiesaccruing to
other firms.
Employees
with
specific
raininghave
less
incentive o quit,
and firms ave less
incentive to
fire
them, than employees
with no
or
general training,
which
im-
plies that
quit
and
layoff ateswould be
inversely related
to the
amount of
specific raining. urnover
would be least
for
employees
with
extremely specific
training
and most
for those
receiving
such general trainingthat productivity
was raised
less
in
firms
providing the
training
than elsewhere.These
proposi-
tions are as
applicable
to
the
large
amount
of
irregular
quits
and
layoffs
that
continually
occur
as to
the more
regular cyclical and secular
movements
in
turnover;
in
this
section, however,
only
the
more
regular
movements
are
discussed.
Consider a
firm
that experiences an
unexpected
decline
in
demand for
its
-
8/17/2019 Investigación en Capital Humano Becker
15/42
22
GARY S. BECKER
output,
the rest
of the
economy
being
unaffected.
The marginal
product
of
employees
without
specific
training-
such
as untrained
or generally
trained
employees presumablynitially
qualed
wages,
and
their employment
would
be
reduced
to prevent
theirmarginal
pro-
ductivity
from
falling
below wages.
The
marginalproduct
of
specifically
rained
employees
initially
would
have been
greater
han
wages.
A decline
n demand
would
reduce these marginal
products
too,
but
as
long
as they
were reduced
by
less than
the
nitial
difference
ithwages,
firmshave no incentiveto lay off uch
employees.
For sunk
costs are sunk,
and
there
s no
incentive
o lay
off
mployees
whose
marginalproduct
s
greater
than
wages,
no
matter
how
unwise
t
was,
in
retrospect,
o invest
in
their training.
Thus
workerswith specific
raining
eem
less likely
to
be
laid
off
s a consequence
of a
decline
n
demand
than
are
untrained
or even
generally
rained
workers.'7
If the decline in demand were suf-
ficiently reat
so
that
even
themarginal
product
of specifically
rained workers
was pushed
below
wages,
would the
firm
just
proceed
to
lay
them off until
the
marginal
product
was brought
into
equality
withwages?
To
show the
danger
here, ssume
thatall
the
cost and
return
from pecific
raining
was
paid and
col-
lected by
the
firm.Any worker
aid off
wouldtry ofind new ob, sincenothing
would
bind
him
to
the old
one.'8
The
firmmight
be
hurt
if
a
new
job
was
found,
for the firm's nvestment
n his
17
A very
imilar rgument
s developedby
Wal-
ter Oi
in "Labor as a Quasi-fixed
actorof
Produc-
tion" unpublished
h.D.
dissertation, niversity
f
Chicago).
18
Actually
one need only
assume
that the quit
rate of laid-off
workerstends
to
be significantly
greater han that ofemployedworkers,f onlybe-
cause
the
cost of
searching
or
nother ob
is less
for
laid-off
orkers.
training
might
be lost
forever.
f
spe-
cifically
trainedworkers
were
not
laid
off, the
firm
would
lose
now because
marginal product
would
be less
than
wagesbutwouldgain
in
the
futuref the
decline
in demand proved
temporary.
There is an
incentive,
herefore,
ot
to
lay
offworkers
with
specific
training
when
their marginal
product
is only
temporarily
elow
wages,
and
the larger
a firm's
nvestment
he
greater
the
in-
centive
not to
lay
off uch
workers.
A worker
ollecting
ome of thereturn
from pecific
raining
wouldhave
less
in-
centiveto finda new job when tempo-
rarily
aid off han
others
would:
he does
not want
to
lose
his investment.
His
be-
havior
while laid off
n turn
affects
his
chances
of
being
laid off,
for
if
it
were
known
that
he would not
readily take
another
ob,
the firm
ould
lay
him
off
without
much
fear of losing
its
invest-
ment.
The
conclusion
here can
be
briefly
summarized.When one firmalone ex-
periences
an unexpected
decline in de-
mand,
relatively
few
workerswith
spe-
cific
training
would
be
laid
off,
f
only
because
their
marginal
product
were
initially reater
hantheir
wage. f the
de-
cline
were
permanent,
ll workerswould
be
laid
off
when
their
marginal
product
became
less than
their
wage and
all those
laid off
would
have
to find
jobs
else-
where. If the decline were temporary,
specifically
rained
workers
might
not
be
laid off
even
though
their
marginal
product
were less than
their
wage
be-
cause
the
firm
would
suffer
f
they
took
other
obs.
The
likelihood
f their
aking
other
jobs
would
be
inversely
related,
and
therefore
the likelihood of
their
being
aid
off
would
be
positively
elated,
to
the
extent
of
theirown investment
n
training.
The analysiscan easily
be
extended
to
-
8/17/2019 Investigación en Capital Humano Becker
16/42
INVESTMENT
IN HUMAN
CAPITAL:
A THEORETICAL
ANALYSIS
23
cover general
declines in demand;
sup-
pose,
for
example, a
general
cyclicalde-
cline
occurred. et me assume
that wages
are sticky nd remain
t theinitial evel.
If the decline in business activity
were
not
sufficient o reduce
the
marginal
product below
the
wage, workerswith
specific
training would not
be laid off
even
thoughothers
would be, just as
be-
fore.
If
the
decline
reduced marginal
product
below wages,
only one modifica-
tion
in
the previous
analysis
is required.
A
firm
would
have a greater
ncentive
o
lay
off pecifically rained
workers
han
when it alone experiences decline be-
cause
laid-off orkerswould
be
less
ikely
to
find other obs
when
unemployment
was widespread.
In other respects
the
implications
of a general
decline with
wage
rigidity
re the same as those
of a
decline
n
one
firm
lone.
The
discussion has concentrated
on
layoff
ates,but the same
kind of reason-
ing
shows that a rise
n wages elsewhere
wouldcause fewer uitsamongspecifical-
ly
trained
workers
than among
others.
For
specifically
rainedworkers
nitially
receivehigher
wages than are
available
elsewhere and
the
wage
rise
elsewhere
would
have to be
greater
han
the
initial
difference efore they would
consider
quitting.
Thus both the
quit
and
layoff
rateof specifically
rained
workers
would
be
relatively
ow and
fluctuate
elatively
lessduring usinesscycles.These are im-
portant
mplications
han can
be tested
with the data available.
Although
quits
and
layoffs
re
influ-
enced
by
considerations ther
than
in-
vestment
osts,
ome
of
these,
uch
as
the
presence
of pension plans,
are
more
strongly
elated
o
investments
han
may
appear
at
first
blush.
A
pension
plan
with
incomplete
vesting privileges'9
penalizes employeesquittingbeforere-
tirement
nd thus
provides
an
incentive
-often an extremely
owerful ne not
to quit.
At the same timepension plans
"insure" firms gainst
quits for hey are
given a lump sum-the
non-vestedpor-
tion of payments-whenever a worker
quits.
Insurance s neededfor pecifically
trained mployees ecause
their urnover
would impose capital
losses on firms.
Firmscan discourage uch
quits by shar-
ing training costs and
the return with
employees,
but
they have less need to
discourage
them
and
would be more
willing
o pay
for
training osts if nsur-
ance
was provided.
The
effects
n the
n-
centive o invest n one's employeesmay
have been
a
major
stimulus to the de-
velopment
of
pension plans.20
An effectiveong-term
ontractwould
insure firms
gainst
quits, just as pen-
sions
do,
and also insure
employees
against layoffs.
Firms
would
be more
willing o pay
for ll kinds
of
training-
assumingfuturewageswere set
at an ap-
propriate level-since a
contract,
in
effect, onverts all training into com-
pletely pecific raining.
A
casual reading
of
history uggests
that
long-term
on-
tracts
have, indeed, primarily
been
a
means of
inducing
firms
to
undertake
large
investments
n
employees.
These
contractsare
seldom
used
today
in
the
United
States,2'
nd while
they
have
de-
clined in
importance
over
time, they
were
probably always
the
exception
here
largely because courts have considered
them
a form of
involuntary
servitude.
19
According
o the National Bureau
ofEconomic
Research tudy
ofpensions,mostplansstillhave in-
completevesting
see
D.
Holland's
report n A Re-
spectfor Facts:
National
Bureau
of Economic
Re-
searchAnnual
Report New York: National Bureau
of Economic Research,
1960],pp. 44-46).
20
In
recentyears
pensionshave also been an
im-
portant ax-saving evice, which ertainly
as been
a crucialfactorn theirmushrooming rowth.
21
The military nd entertainment
ndustry re
the major exceptions.
-
8/17/2019 Investigación en Capital Humano Becker
17/42
24
GARY S.
BECKER
Moreover,
ny enforcible
ontractcould
at
best specify he
hours requiredon a
job,
not the
quality
of performance.
Since performance
an
varywidely, un-
happyworkers ould usually "sabotage"
operations o
induce
employers o release
them
from ontracts.
Some trainingmay
be useful
neither
in most
nor only
n a singlefirm
ut in a
set of
firms efinedby
product,
type of
work,
or geographical
ocation.
For
ex-
ample, carpentry
training
would raise
productivity
primarily
n
the construc-
tion industry,
nd
French
egal training
wouldbe ineffectiventhe UnitedStates,
with
its differentanguage
and
legal
institutions.
Such training
would tend
to be paid by trainees,
ince
a single
firm
couldnot readily
ollect
he
return,22
nd
in this
respectwould be
the same
as gen-
eral training.
n one respect,however,
t
is similar
to specific training.
Workers
with
training"specific"
to
an
industry,
occupation,
or country
re less likely to
leave thatindustry, ccupation,or coun-
try (via migration)
than other workers,
so their industrial,
occupational,
or
country
"turnover"
would
be less than
average.
The same
result s obtained
for
specific
training, except
that a
firm
rather
than an industry,
ccupation,or
country
s
used as
the
unit
of
observa-
tion
in
measuring
urnover.
An
analysis
of
specific
raining, herefore,
s
helpful
also in understanding he effects f cer-
tain
types
of
"general"
training.
Although discrepancy
etween
mar-
ginal product
and wages
is
frequently
takenas evidence
of mperfections
n
the
competitive
ystem,
t
would occur
even
in a perfectly
ompetitive nvironment
where
there
is investment
in
specific
22
Sometimes irms o-operate
n
payingtraining
costs, especiallywhen training pprentices see A
Look at Industrial
Training n MercerCounty,
N.J.
[Washington
ureau
of
Apprenticeship
nd
Train-
ing,
19591,
. 3).
training.
The investment
pproach
pro-
vides
a very
different
nterpretation
f
some
commonphenomena,
s canbe seen
from
he following
xamples.
A positive difference etween
mar-
ginal product
and
wages
is
usually said
to
be
evidence
ofmonopsony ower,
nd
just
as the ratio
of
product
price to
marginal
cost has been
suggested
as a
measure
ofmonopolypower,
so has the
ratio
of marginalproduct
to wages been
suggested
as a
measure
of
monopsony
power.
But
specific
rainingwould
also
make
this ratio greater
than
one. Does
the differencebetween the marginal
product and
the earnings
of major-
league
baseball players,
for example,
measure
monopsonypower
or
the return
on a team's investment? ince
teams do
spend
a great
deal on developingplayers,
some
and
perhaps
most of
the
difference
must
be considereda
returnon invest-
ment even
were there
no
uncertainty
about
the
abilities
of
different
layers.23
Earningsmightdiffer reatlyamong
firms,