document resume ed 089 396 24 81 005 983document resume 24 81 005 983 innes, jon t.; and others the...

ED 089 396

AUTHORTITLE

INSTITUTION

SPONS AGENCY

BUREAU NOPUB DATECONTRACTNOTEAVAILABLE FROM

EDRS PRICEDESCRIPTORS

DOCUMENT RESUME

24 81 005 983

Innes, Jon T.; And OthersThe Economic Returns to Education. A Survey of theFindings. CASEA Monograph No. 3.Oregon Univ., Eugene. Center for Advanced Study ofEducational Administration.National Center for Educational Research andDevelopment (DHEN/OE), Washington, D.C.BR-5-0217-6May 65OEC-4-10-16353p.; A re-announcement of ED 011 572Publications, Center for Educational Policy andManagement, University of Oregon, Eugene, Oregcn97403 ($1.00)

MF-$0.75 HC-$3.15 PLUS POSTAGECollege Graduates; *Economic Research; *EducationalBenefits; *Educational Economics; Higher Education;High School Graduates; *Human Resources; *Income;Literature Reviews; *School Support; School Taxes

ABSTRACTThis monograph is a survey of recent literature on

the economic returus on investment in education. The first sectionintroduces the reader to the concept of human capital and to someways it can be measured. In the second and third sections, data arepresented on the relationship between education and income and thebenefits accruing both to individuals and to the nation as a wholefrom investment in education. Section four deals with estimates ofeducation's contribution to economic growth. Section five summarizesthe generalizations concerning education as an investment. Thesestudies reinforce the conclusion presented by other studies that thesocial returns on educational investment are very profitable, as areprivate returns on individual investment. A 32-item bibliography isincluded. (TC)

cri4,

THEuicz

ECONOMIC RETURNSTO EDUCATIONA Survey of the Findings

JON T. INNL:S

PAUL B. JACOBSON

ROLAND J. PELLEGRIN

9 S DEPARTMENT OF HEALTH.EDUCATION& YSELFARENATIONAL INSTITUTE OF

EDUCATIONTHIS OCCQMENT HAS BEEN REPRODUCE° EXACTLY AS RECEIVED FROMTHE PERSON OR ORGANIZATION ORIGINAt11,40 IT POINTS OF VIEW OR OPINIONSSTATED C4 NOT NECESSARILY FEPAESENT OF FiC,AL NATIONAL INSTITUTE OFEDUCATION POSITION OR POLiCY

A Publication of

THE CENTER FOR THE ADVANCED STUDY OFEDUCATIONAL ADMINISTRATION

UNIVERSITY OF OREGON

Eugene, Oregon

t.

First Printing, May, 1965Second Printing, Febraoy, 1966Third Printing, N ovember, 1969

9)Printed in the United States of America

at the University of Oregon Press, Eugene, Oregon

Designed and edited by Joanne M. Kitchel

The text of this book is set in Janson typeon Simpson's eggshell text

Library of Congress Catalog Number 66-63896

;1.

The Ccntcr for the Advanced Study of Educational Adminis-tration at the University of Oregon is a national research anddevelopment center which is supported in part by funds fromthe United States Office of Education, Department of Health,Education, and Welfare. The opinions expressed in this publi-cation do not necessarily reflect the position or policy of theOffice of Education and no official endorsement by the Officeof Education should be inferred.

INTRODUCTIONDuring the past few years economists have done highly

significant research on the economic benefits accruing to theindividual and to society from investment in education. Thisresearch has found, as many people would expect, that aninvestment in high school or college education is quite profit-able for the student. Less known, and perhaps more surpris-ing, is the conclusion of researchers that investment ineducation contributes to the public welfare, or to the eco-nomic well-being of the nation as a whole.

Perhaps this finding surprises many of us because we areunaccustomed to thinking of education as an "income pro-ducer"; rather, most of us think of schools and colleges as"tax eaters." The fact of the matter is, however, that econ-omists have found a high "public" return on investment ineducation. They have also concluded that education makesan important contribution to the economic growth of ournation.

This monograph is a survey of recent and useful literatureon the economic returns to education. We have attempted tosummarize basic findings. Readers who wish more detailedinformation are, of course, invited to consult our sources,which are listed in the bibliography at the end of thismonograph.

While our presentation deals solely with the economic re-turns to education, we are well aware that education mayconfer numerous other benefits upon the individual and mayhave many other kinds of consequences for the country thaneconomic ones alone. We also recognize that a number ofmatters other than education affect the earnings of the indi-vidual. Renshaw (20) has listed the following factors as beingpositively correlated with formal education. Each undoubt-edly contributes to a person's earning power.

(1) People with high IQ's generally obtain more educa-tion.

[v]

(2) People with more education work longer hours.

(3) Other kinds of education are closely related to theamount of formal education one obtains.

a. Self educationb. Experiencec. Education obtained in the homed. Training in the armed forcese. On-the-job training

(4) Some fields that require a high degree of specializationhave restrictions on entry.

These and other factors point to the limitations of ourapproach. If we attribute the entire increase in earning powerto formal education, we are no doubt overstating our case.

The reader should also note that much of the analysis thatfollows uses the rate-of-return approach. As stated by Hunt(16), there are essentially four assumptions which have beenmade, either implicity or explicitly, by those who have esti-mated rates of return to education. Each of these could obvi-ously introduce biases. These assumptions are

(1) Private product is a satisfactory representation of so-cial product.

(2) Rates of return on physical and educational capital areconceptually similar.

(3) Analysis of cross section data provides useful estimatesfor projecting trends into the future.

(4) The associated increase in money income is a satisfac-tory measure of the private returns to education.

To the extent that any or all of these are unjustified, thenclearly any policy implications drawn from the results arealso unjustifie0.

The organization of the monograph is as follows. The firstsection huroduces the reader to the concept of "human capi-

I vi

tal" and to some ways it can be measured. In the second andthird sections, data are presented on the relationship betweeneducation and income and on the "public" and "private" ratesof return on investment in education. Section four deals withestimates of education's contribution to economic growth,and section five summarizes the generalizations drawn In themonograph.

CONTENTS

INTRODUCTION V

IThe Concept and Measurement of Human Capital

II

Education and Earnings

111The Rate of Return on Investment in Education 18

IVThe Contribution of Education to Economic Growth 33

VSurmnary

BIBLIOGRAPHY

39

I

The Concept and Measurement ofHuman Capital

EDUCATION As AN INVESTMENT IN HUMAN CAPITAL

The essential characteristic of education which precludesour thinking of it solely as a consumption item is that thebenefits which accrue from education do so over a longperiod of time. While it is no doubt true that there are sub-stantial immediate benefits of education which give it theappearance of being a consumption item, study of Table 1indicates that the long-term benefits represented by the in-come stream are considerable.

What Table 1 tells us is that, on the average, the annualincome of males in a given age bracket increases as educationincreases. Thus, even at age 65, over 40 years after his formaieducation ceased, the average college graduate earned $2,266more in 1949 than his counterpart whose formal educationended after four years of high school. While it would bewrong to attribute the entire amount of the differences shownin Table 1 solely to differences in levels of education, itwould be equally wrong to contend that they result entirelyfrom differences in ability, motivation, family position orother such factors which have a bearing on the amount of anindividual's income. The fact that the differentials have per-sisted in spite of the increasing number of people now pursu-ing higher education does indicate, moreover, that educationitself is a dominant factor (see Miller, 17).

By treating education as an investment in human capital,we are saying that the process of education enhances theproductivity of an individual, and that this increase in pro-ductivity is reflected in the income streams as shown in Table1. We are by no means suggesting that monetary gain is the

Tab

le I

AV

ER

AG

E I

NC

.OM

E B

Y A

cE A

ND

YE

AR

S O

F SC

HO

OL

CO

MPL

ET

ED

MA

LE

S, U

NIT

ED

ST

AT

ES,

194

9

Agc

01-

4

Yea

rs o

f Sc

hool

Com

plet

edE

lem

enta

ry S

choo

lH

igh

Scho

ol5-

78

1-3

4C

one

1-3

4+14

-15

$ 61

05

350

$ 36

5$

406

16-1

752

647

251

4S3

4$

429

18-1

968

471

388

51,

069

941

$ 95

520

-21

944

1,00

91,

216

1,53

51,

652

1,74

4$1

,066

22-2

41,

093

1,22

71,

562

1,93

12,

191

2,36

31,

784

$1,9

2625

-34

1,33

71,

603

2,02

72.

540

2,83

73,

246

3,44

44,

122

35-4

41,

605

1,84

22,

457

3,02

93,

449

4,05

55,

014

7,08

545

-54

1,81

22,

073

2,65

03,

247

3,72

54,

689

5.63

98,

116

55-6

42,

000

2,04

52,

478

3,01

03,

496

4,54

85,

162

7,65

565

or

mor

e1,

140

1,18

91,

560

1,89

82,

379

3,15

53,

435

5.42

1

Sour

ce: H

anse

n (1

4). T

able

1, 1

30.

The Concept and Aleaptrement of tiuntan Capital

only motivation of those who seek higher education; it obvi-ously is not. What we arc suggesting is that an individualwho is considering an expenditure now which will yieldfuture income ought to consider education as one of thealternatives.

Thus far, we have been considering human capital fromthe individual's point of view. It is also important to considerit from society's point of view. Thus, we might ask the ques-tion: N1'h it would be the economic effect of increased publicexpenditure on education? While the answer to such a ques-tion would depend on the amount and distribution of suchexpenditure, there is reason to believe that education has beenan important determinant of the rate of growth of output inthis country. According to the figures in Table 2, between1919 and 1957 output increased at an average annual rate of3.1 per cent, while inputs of labor and physical capital in-creased by only 1.0 per cent per year. Assuming no signifi-cant economics of scale, this leaves about two-thirds of thegrowth in output unaccounted for by changes in inputs. Thehypothesis which immediately suggests itself is that at leastsome of this growth is a result of investment in human capital.

Table 2INcREAses IN OUTPUT A1.41) INPUTS OF THE PRIVATE DOMESTIC

SECrOR OF THE LharED STATES ECONOMY, 1919 TO 1957

1919 1957 Increases in Per Cent(Indexes 1929e-400) Per Annum

Output . 69.7 225.2 3.1Input

Labor Input ... 86.7 116.9 0.8Capital Input ..... ..... 80.3 158.2 1.8Total (weighted by

relative shares) 84.9 125.5 1.0

Source: Schultz (22), Table I, SO.

THE MEASURVMENT OF HUMAN CAPITAL

Because the amount of education possessed by an individualis generally measured in terms of years, it is quite natural to

(3)

THE ECONOMIC RETURNS TO EDUCATON

suggest such a unit of measurement for the total stock ofhuman capital represented by the aggregate education ofthe labor force. Thus, it is tempting to say that if the averagemember of the labor force has had 11 years of school and thetotal labor force is 70 million persons, the total stock of hu-man capital is equivalent to 770 million "school years of edu-cation." The problem with such a measure is that it fails totake into account such things as differences in the extent towhich various years of school contribute to productivity,differences in types of education (e.g., general versus tech-nical), changes in the number of days of school attendance,and long-run changes in the quality of education.

Schultz has developed a more sophisticated approachwhich, while still containing many of the same deficienciesmentioned above, is superior to the "aggregate education"measurement. This procedure involves measuring the stockof education in terms of "equivalent 1940 school years," ameasurement which takes into account the effects of changesin school attendance. His results, given in Table 3, show thatthe stock of human capital increased three and one-half foldbetween 1900 and 1957 when measured in terms of schoolyears, but increased six and one-third fold when measuredin "equivalent school years."

!41

Tab

le 3

Tom

SC

HO

OL

YE

AS

O C

ompr

zav

or x

xxe

LAno

tt F

un=

or

TH

E T

JNrn

m S

rAnt

s, 1

900

2+01

957

Equ

iv. 1

940

Labo

rS

choo

l Yea

rsf T

otal

Sch

ool

Sch

ool Y

ears

For

ceC

ompl

eted

Yea

rs C

ompl

eted

Com

plet

edY

ear

(=M

ons)

Per

Per

son

(mill

ions

)P

er P

erso

n...

.%

' 190

028

.17.

7021

64.

14...

1910

35.8

7.91

283

4.65

1920

41.4

8.12

336

5.25

1930

48.7

8.41

410

6.01

1940

52.8

9.02

476

7.24

1950

60.1

10.1

060

78.

65

1957

70.8

10.9

677

610

.45

Tot

al E

quiv

..19

40 S

choo

lYea

tsC

ompl

etec

l(=

Zoo

s)

116

167

217

293

382.

520

740

Sou

rce:

. Sch

ultz

(22

). T

able

1.)

,


As suggested earlier, the extent to which productivity isrelated to education can be seen by assuming that earningsare an accurate reflection of productivity and comparingthe earnings streams of individuals with various mnounts ofeducation. In this section we shall make three types of com-parisons:

(I) Annual incomes of persons in the same age bracketwith different amounts of education.

(2) Lifetime incomes of individuals with different amountsof education.

(3) Discounted lifetime incomes of individuals with differ-ent amounts of education.

EDUCATION AND ANNUAL INCOME

Table 4, taken from Miller (17), shows variations in aver-age annual income over the past generation for males withdifferent amounts of education. Women are excluded fromthe analysis because the relationship between their incomeand education may be distorted by the fact that a large pro-portion of them either do not enter the labor force or areemployed only on a part-time basis.

The data In Table 4 clearly show that at all age levels addi-tional schooling is associated with higher average incomesfor males. This association has persisted despite the fact thatthe educational attainment of the population has increasedconsiderably in the past generation. In fact, as shown in Table5, the differential between high school graduates and elemen-tary school graduates has widened, while that between col-lege graduates and high school graduates has shown little

[6]

Tab

le 4

Mr-

Arr

IN

couc

z (o

n E

mon

ams)

Foe

MA

L 2

5 Y

EA

RS

or A

a A

ND

Ova

HY

YE

AR

S or

SC

HO

OL

CO

MPL

EC

ED

AN

D A

m 1

939,

194

6,19

49, 1

956,

AN

D 1

958

Yea

rs o

f Sc

hool

Com

plet

edan

d A

ge19

3919

46b

1949

e19

56e

1958

e

Tot

al: 2

5 Y

ears

Old

and

Ove

rE

lem

enta

ry: T

o.4

7403

6$2

,041

$2,3

94$3

307

$3,0

96L

ess

than

8 y

cars

d(e

)1.

738

2,06

22,

613

2,55

18

year

s(e

)2,

327

2,82

930

23.

769

Hig

h Sc

hool

: 1 to

3 y

ears

1,57

92,

449

3,22

64,

480

4,61

84

year

s1,

661

2,93

93,

784

5,43

95,

567

Col

lege

:1

to 3

yea

rsL

951

3,65

44,

423

6363

6,96

64

year

s or

mor

e2,

607

4,52

76,

179

8,49

09,

206

25 to

34

Yea

rs:

Ele

men

tary

: Tot

al83

71,

729

2,18

53,

061

3.14

3L

ess

than

8 y

ears

d(e

)1,

394

1,88

02,

662

2,67

08

year

s(e

)2,

011

2,54

03,

685

3,66

3H

igh

Scho

ol: 1

to 3

yea

rs1,

150

2,06

22,

837

4,40

74,

341

4 ye

ars

1,33

52,

335

3,24

64,

813

4,90

9C

olle

ge:

1 to

3 y

ears

1,56

62.

875

3,44

45,

437

5,77

44

year

; or

mor

e1,

956

3,23

74,

122

6,30

77,

152

tn,

Tab

le 4

- C

ontin

ued

Yea

rs o

f Sc

hool

Com

plet

edan

d A

ge19

319

46b

1949

e19

56e

1958

e

75 to

44

Yea

rs:

Ele

men

tary

: Tot

al1,

110

2095

2.61

03,

694

3,68

6L

ess

than

8 y

ears

d(e

)1,

730

2,24

43.

169

3,02

38

year

s(e

)2,

425

3,02

94,

256

4,40

3H

igh

Scho

ol: 1

to 3

yea

rs2,

574

2,60

73,

449

4,79

95.

035

4 ye

ars

1,97

93,

463

4,05

559

926.

007

Col

lege

:1

to 3

yea

rs2,

270

4,06

95,

014

7.13

18,

015

4 ye

ars

or m

ore

3,14

15,

054

7,08

59,

790

10,1

06

45 to

S4

Yea

rs:

Ele

men

tary

: Toa

d1.

199

2,34

92,

797

3.67

23,

660

Les

s th

an 8

yea

rs4

(e)

2,02

72,

418

5.07

83,

008

8 ye

ars

(e)

2,62

93,

247

4,28

943

37H

igh

Scho

ol: 1

to 3

yea

rs1,

732

2,95

93,

725

4,87

64,

864

4 ye

ars

2,25

63,

744

4,68

96.

104

6,29

5C

olle

ge:

1 to

3 y

ears

2,42

84,

671

5.63

97,

426

8,68

24

year

s or

mor

e3,

575

5,24

28,

116

11,7

0212

,269

55 to

64

Yea

rs:

Ele

rnm

atar

y:T

otal

1,05

72,

082

2,57

73,

462

3,43

6L

ess

than

8 y

ears

(e)

1,81

42,

278

2922

2,95

68

year

s(e

).3

653,

010

3,93

23,

960

THE ECONOMIC RETURNS TO EDUCATION

change, This latter differential did, however, increase con-siderably during the recession years of 1949 and 1958, suggest-ing the possibility that persons with college educations arenot as subject to unemployment as are the less educated.

Table 4 also shows that, in terms of income received, col-lege-trained individuals seem to benefit the most from yearsof experience on the job. Table 6 gives further informationon this matter. In 1958, college-trained Individuals in the 45to 54 age bracket (peak earning power years) had mean in-comes which were 72 per cent higher than the mean incomesof college-trained individuals who were, on the average, 21)years younger, The comparable percentages for high schooland elementary school graduates were 28 and 18, respectively.Thus, the colkee graduate of any given age not only earnsmore than persons of the same age with less education, butthe earnings of the college graduate exceed those of otherpersons to a greater extent as he grows older. In other words,the college graduate is increasingly advantaged with respectto earnings as he grows older.

EDUCATION AND LIFETIME INCOME

In attempting to measure lifetime income it would, ofcourse, be ideal to have life-cycle data on a group of indivi-duals. In the absence of such data, Miller has constructedsome estimates based on the data shown in Table 4, adjustedfor mortality rates. These figures, which are shown in Table7, arc computed by summing the earnings at each age multi-plied by the probability that an individual will live to thatage. Using the figures for 1949, we find support for the fre-quently heard remark that a college education is "worth"$100,000. That is, the difference in lifetime income of acollege graduate and a high school graduate was expectedto be, in 1949, about $100,000 (more accurately, it was over$111,000). By 1958, this difference in lifetime income had in-creased to approximately $178,000.

As Miller has indicated, the more highly educated groups

[ 101

Tab

le 5

Mw

bec

omae

(oa

EA

R:0

:ms)

BY

Lem

or

Savi

or. C

ompr

i:1m

, FO

IL M

AL

IM 2

5 Y

xrut

s O

LD

AN

D O

VE

R, F

OR

TH

EU

mrE

D &

mum

: 193

9, 1

996,

194

9, 1

956,

AN

D 1

958

Ele

men

tary

-H

igh

Scho

ol D

iffe

rent

ial

Hig

h Sc

hool

-Col

lege

Dif

fere

ntia

lA

vera

ge I

ncom

eA

vera

ge I

ncom

e

Ele

men

tary

Hig

hH

igh

Scho

olSc

hool

Per

Cen

tSc

hool

Col

lege

Per

Cen

tY

ear

Gra

duat

eG

radu

ate

Dif

fere

nce

Gra

duat

eG

radu

ate

Dif

fere

nce

1939

(a)

$1,6

61(a

)$1

,661

52,6

0757

1946

52.3

272,

939

2629

394,

527

5419

492,

829

3,78

434

3,78

46,

179

6319

563,

732

5,43

946

5,43

98,

490

5619

583,

769

5,56

748

5,56

79,

206

65

Not

ava

ilabl

e.So

urce

: Mill

er (

17),

Tab

le 3

, 969

-

Tab

le 6

Mr-

lor

larc

orez

Oa

EA

RN

ING

S)

FO

R M

Ats

s 25

TO

34

YE

AR

S A

rm 4

5 T

o 54

YE

AR

S o

r A

cr.,

zet L

am. o

rS

cuoo

t. C

om

,B

oa T

irg U

Nriz

u S

TA

TE

S: 1

939,

194

6. 1

949,

195

6, a

nd 1

958

Age

and

Lev

el o

fS

choo

l Com

plet

ed19

3919

4619

4919

5619

58

Ele

men

tary

Sch

ool G

radu

ate

25 to

34

year

s(a

)$2

,011

$2,5

40$3

,685

$366

3

45 to

54

year

s(a

)2,

629

3,24

74,

289

4,33

7

Per

Cen

t Inc

reas

e(a

)31

2816

18

Hig

h S

choo

l Gra

duat

e25

to 3

4 ye

ars

51,3

3523

353,

246

4,81

34,

909

45 to

54

year

s2,

256

3,74

44,

689

6,10

46,

295

Per

Cen

t LIC

XC

ISC

6960

4427

28

Col

lege

Gra

duat

e25

to 3

4 ye

ars

1,95

63,

237

4,12

26,

307

7,15

2

45 to

54

year

s3,

575

5,24

28,

116

11,7

0212

269

Per

Cen

t Inc

reas

e83

6297

8672

(a)

Not

ava

ilabl

e.S

ourc

e: M

iller

(17

), T

able

6.9

74.

Tab

le 7

Dre

rim

e Im

mix

(E

voro

zcs)

BA

SED

ow

Aar

rnaa

Eri

c IV

IzA

rrs

roe

MA

L=

nv

Sara

n A

ce G

RO

UPS

. BY

YrA

rs o

r Sc

Em

ot.

CO

MPL

EIE

D, F

OR

ME

Min

im S

urzs

: 193

9, 1

946,

194

9,19

54 A

ND

195

8

Yea

rs o

f Sc

hool

Com

plet

ed a

nd A

ge19

3919

46b

1949

e29

56.7

1958

c

Inco

me

from

Age

18

to D

eath

:E

lem

enta

ry: T

otal

(e)

(e)

$113

,330

$154

.593

$154

.114

Les

s th

an 8

yea

rsd

(e)

(e)

98.2

2213

2,73

612

9.76

48

year

s(e

)(e

)13

2.68

318

0,85

718

1,69

5H

igh

Scho

ol: 1

to 3

yea

rs(e

)(e

)15

2.06

820

5,27

721

1,19

34y

ears

(e)

(e)

185,

279

253,

631

257,

557

r C

olle

ge:

I to

3 y

ears

(e)

(e)

2092

8229

/581

315,

504.

4yea

rs o

r m

ore

(e)

296,

377

405,

698

435,

242

%D

J

Inco

me

from

Age

25

to D

eath

:E

lem

enta

ry: T

otal

(e)

$ 87

,004

104,

998

143,

712

143,

808

Les

s th

an 8

yea

rsd

(e)

74,3

6991

,095

123,

295

120.

965

8 ye

ars

(e)

98,7

0212

2,78

716

8,00

416

9,97

6H

igh

Scho

ol: 1

to 3

yea

rs(e

):u

7,94

014

1,87

019

2,25

419

8,88

14y

ears

(e)

135,

852

174,

740

237,

776

241,

844

Col

lege

:1

to 3

yea

rs(e

)16

1,69

920

1,93

828

1,55

330

5,39

54y

ears

or

mor

e(e

)20

1.73

128

6,83

339

1992

419,

871

Res

tric

ted

to p

erso

ns r

epor

ting

$1 o

r m

ore

of w

age

or s

alar

y in

com

ean

d le

ss th

an $

50 o

f ot

her

inco

me

for

nativ

e w

hite

s an

d N

egro

es.

b T

otal

mon

ey e

arni

ngs.

year

s of

sch

ool c

ompl

eted

, not

sho

wn

e T

otal

mon

ey in

com

e.se

para

tely

.In

clud

es p

erso

ns r

epor

ting

noN

ot a

vaila

ble.

Sour

ce: A

IM=

(17

), T

able

11,

981

.


have made the greatest relative gains in lifetime Income inthe years since 1939. For example, in 1946 high school gradu-ates could expect to earn 35 per cent more than elementaryschool graduates; by 1958, however, they could expect toearn 44 per cent more. The income of high school graduateshas thus been rising more rapidly than that of persons withless education. The same pattern, but in pronounced form,exists between the earnings of high school and college gradu-ates. For instance, the differential in favor of college gradu-ates was 48 per cent in 1946. By 1958, it had risen to 70 percent (Miller, 17, 983).

EDUCATION AND DISCOUNTED LIFETIME INCOME

It is probably safe to assume that a rational individual,when confronted with the problem of measuring his lifetimeincome, would be concerned not only with its amount butalso with its distribution throughout his future. In fact, hewould tend to value a given amount of income more highlyif it were to be received in the relatively near future as op-posed to the distant future, and he would value it most highlyif he were to receive it as current income. This being thecase, he should have some method of reducing a given amountof future income to its equivalent in current income. Thediscounting procedure, to be examined here, does just that.

By way of example let us assume that an individual is ex-pecting to earn $5,300 in the year following the current year.Let us further assume that he could invest his money, if he sodesired, and earn interest at the rate of 6 per cent per year.The "present value" of the $5,300 is therefore, the amountwhich he would have to invest now so that in one year's timeit would appreciate to $5,300. Let us call this amount C,.Clearly, then, it must be the case that

CI (1.06) = $5,300therefore

$5,300= = $5,000

1.06

[ 14 j


That is, the present value of the $5,300, to be received oneyear hence, is $5,000. Likewise, if our hypothetical incomeearner is expecting to earn $6,000 in the second year, the"present value" of this amount, call It C2, can be found fromthe expression:

or,

Ci (1.06)2 = $6,000

$6,000

C, = ;5,340.86(1.06)2

In general, then, if we designate income expected yearshence as Et, we may calculate the present value of E,, call it

by the formula

C,(1 + r)2

where r is the assumed rate of interest.

The present value of the entire stream of lifetime earningsis thus the sum of all the individual Ci's. In effect this pro-cedure provides an answer to the question of how much anindividual would have to invest currently, at a given interestrate, in order to receive the same stream of future receiptswhich he now expects as earnings.

Houthakker (15), using the same basic data as Miller, hascomputed mean incomes in 1949 by age and years of school-ing both before and after taxes. He then weighted these fig-ures to reflect mortality rates, thus providing estimates ofexpected lifetime incomes for persons with various levels ofeducation. Using the techniques described above he has thendiscounted these lifetime incomes back to age 14 using dis-count rates of 0, 3, 6 and 8 per cent. The results of this pro-cedure are shown in Table 8.


Table 8PRESENT VALUE AT AGE 14 OF LIFE-TIME INCOME sv

YEARS or Scitoot. Comearto

Discount Rate(Per Cent) 0 3 6 8

Before Tax

Years of SchoolCompleted

Elementary:0 $ 64,132 $ 26,220 $13,014 $ 8,8961.4 79,386 33,939 17,492 12,1795.7 100,430 42,758 21,834 15,0988 124,105 52,923 27,037 18,700

High School:1.3 142,522 59,734 30,008 20,5144 175,160 72,475 36,328 24,990

College:1.3 198,268 78,138 36,547 23,7934 or more 280,989 106,269 47,546 30,085

After Tax

Elementary:0 60,785 24,944 12,428 8,5151-4 75,021 32,189 16,638 11,7305.7 93,571 40,006 20,537 14,2528 115,277 49,425 25,380 17,592

High School:1.3 130,933 55,260 27,945 19,1884 157,940 66,055 33,466 23,149

College:1-3 175,206 69,651 32,912 22,4004 or more 238,761 91,335 41,432 26,454

Source: Houthakker (15) Table 3, 28.

For our purposes we are interested in the difference be-tween any two entries in a given column of Table 8. Con-sider, for example, the individual who is contemplating fouryears of college after finishing high school, Using a discountrate of 8 per cent, the present value of the extra earnings, be-

[ 16)

Education and EarnIn s

fore taxes, resulting from a college education is $5,095 (i.e.,$30,085 - $24,990). In other words, the amount of moneywhich a person would have had to invest at age 14 (in 1949)so that he could receive a stream of receipts precisely equalto the stream of extra earnings he would receive because ofhis choice of a college education is $5,095.

In fact, In all but three cases in Table 8, the present valueof the contribution of additional education to the earningsstream is positive. The question now arises as to the extentto which this contribution does or does not exceed the costsof securing the additional education.

It is this question to which we turn our attention in thefollowing section.

17 )

The Rate of Return on Investmentin Education

Thus far, we have been considering only the monetary re-turns to education without making any mention of the costs.Indeed, if we satisfied ourselves with merely summing up thelifetime income stream as above, then costs would appeartrivial and could be neglected. As we have previously noted,as of 1949 a male high school graduate who went on to foiryears of college could expect to add over $100,000 to his life-time earnings (not discounted). This amount is surely muchgreater than the costs of obtaining this education and thusthe investment appears worthwhile.

In this section we will relate monetary returns to educa-tion, as measured in previous sections of this monograph,with costs of education, not yet measured, by means of a dis-count procedure. This procedure provides an answer to thequestion: Whit rate of discount will equate the stream ofextra returns resulting from education with the costs incurredin obtaining that education?

MEASURING THE COSTS OF EDUCATION

We must consider the cost aspect of our problem fromtwo views, social and private. Following Hansen (14), wewill refer to these as "total resource costs" and "private re-source costs." The first, total resource costs, has three majorcomponents: (1) school costs incurred by society (that is,teacher? salaries, supplies, interest, and depreciation on capi-tal); (2) opportunity costs incurred by individuals (namely,income foregone during school attendance); and (3) inci-dental school-related costs incurred by individuals (for ex-ample, books and travel). Private resource costs include num-bers (2) and (3) above, but substitute tuition and fees paid

[18)

The Rate of Return on investment in Education

by individuals for the first kern.Schultz (21) has tabulated total resource costs on an aggre-

gate basis for each of several years between 1900 and 1956.Tables 9 and 10 show a breakdown of the school costs in.cored by society for each of these years. Considering theyear 1950 (which refers to the school year of 1949.50), wesee that the total school costs incurred by society In that yearamounted to 6.505 billion dollars (see Table 9) plus 2.128billion dollars (see Table 10), or a total of 8.633 billion dollars.

Shultz's method of estimating foregone earnings in eachof the years under consideration is as follows. He first findsthe average gross weekly earnings for all manufacturing In.&series. He then climates, on the basis of 1949 data, thathigh school-age workers, being subject to somewhat lowwages, would earn during the school year (40 weeks) anamount equal to 11 times this average weekly wage. College.age workers, according to a similar, estimate, would earn 25times the average gross weekly earnings for all manufactur-ing industries In 40 weeks. Thus, high school students foregothe equivalent of 11 weeks of work at the average manufac.turing wage, and college students forego the equivalent of25 weeks of such a wage. Elementary school students areassumed to have no foregone earnings.

Next Sehult, adjusts his estimate for unemployment. Heassumes the following rates of unemployment and appliesthem to both high school and college students;

YEAR PER CENT UNEMPLOYMENT1000 8.21910 3.91920 4.21930 1241940 14.71950 4.11956 ... . .... . . . ... 3.0

These estimates, both before and after adjustment for unem-ployment, are shown in Table 11,

119]

Tab

le 9

Ikio

rtrA

l. It

moa

ce C

ogs

or E

DU

CA

TIO

NA

L S

ER

VIC

ES

RE

D B

YE

LE

ME

NT

AR

Y A

ND

Sec

oNuA

zy S

lalo

ms

rwrK

E U

Nrc

eu

STA

TE

S, 1

900-

1956

, IN

Cui

szxr

Plu

ms

(Ittr

ions

of

dolla

rs e

xcep

t Col

umn

4 in

Bill

ioos

)

Pnsu

c So

ots

Psiv

AT

E S

CH

OO

LS

Pusr

ac A

mu

Val

ueIm

plic

itPa

w=

Sal

oom

Gro

ssN

etof

Inte

rest

Gra

m

Exp

endi

-C

apita

l Exp

eidi

- Pr

op-

and

De-

Tot

al E

xpea

di-

Tot

alSe

c-E

lem

en-

ture

sO

utla

yta

res

erty

prec

iatio

nPu

blic

ture

sPr

ivat

eT

otal

outla

wta

ry

Yea

r(1

)(2

)(3

)(4

)(5

)(6

)(7

)(8

)(9

)(1

0)°I

>

1900

____

____

_21

535

180

SS44

224

2728

252

1923

3

1910

......

....._

426

7035

61.

188

444

5456

500

SO45

0

1920

1,03

615

488

22.

419

21,

074

104

108

1,18

221

596

7

1930

____

_2,

317

371

1,94

66.

249

62,

442

233

246

2,68

874

11,

947

1940

_2,

344

258

2,08

67.

660

82,

694

227

261

2,95

51,

145

1,81

0

1950

5;83

81,

014

4,82

411

.491

25,

736

783

769

6,50

52,

286

4,21

9

1956

10,9

552,

387

8,56

823

.91,

912

10,4

301,

468

1,40

411

,884

4,03

17,

853

Sour

ce.:

Schu

ltz (

21),

Tab

le 3

, 578

.

The Rate of Return on Investment in Education

f 21 )

Tab

le 1

1A

NN

uAr.

DX

NE

PIG

S F

oarc

oale

sr

Sru

nzie

rs, A

njus

rm A

rm N

O! A

DJU

ST

= M

t tha

MP

LOY

ME

N;

1900

-19%

, be

Cut

untiq

r P

alm

sC

9

Ann

ual E

arni

ngs

For

egon

e P

er S

tude

nt W

hile

Act

oacr

ing

Hig

h S

choo

lC

olle

ge o

r U

nive

rsity

ZA

vera

ge W

eekl

yA

djus

ted

Adi

nste

do X

Ear

ning

s, A

llfo

r U

n-fo

r U

n-5

,--,

Man

ufac

turin

gU

nadj

uste

dem

ploy

men

tU

nadj

uste

dem

ploy

men

tp:

tni

(Dol

lars

)(D

olla

rs)

(Dol

lars

)(D

olla

rs)

(Dol

lars

)Y

ear

(1)

(2)

(3)

(4)

(5)

Z19

0083

792

8420

919

2co

1910

10.7

411

811

326

925

9g

1920

26.1

228

727

565

362

6tx

1 o19

3023

2525

622

458

150

9O

1940

25.2

027

723

663

053

7P

.19

5059

3365

362

61,

483

1,42

2C

1956

80.1

388

185

52,

003

1943

ZS

ourc

e: S

chul

tz (

21),

Tab

le 2

, 575

.

Tab

le 1

2T

OT

AL

RE

SOU

RC

Z C

osrs

roc

Stu

rear

r pr

It Y

r.ou

t or

HIG

H S

CH

OO

LA

ND

CO

LL

EG

E,.

1900

-195

6

HIG

H S

moo

t.C

HIL

D=

Scho

olC

osts

Ear

ning

sFo

rego

neIn

cide

ntal

Cos

tsT

otal

Scho

olC

osts

Ear

ning

sFo

rego

neIn

cide

ntal

Cos

tsT

otal

Yea

r(1

)(2

)(3

)(4

)(5

)(6

)(7

)(8

)

1900

2784

411

517

019

219

381

1910

4511

36

164

228

259

2651

3

1920

8627

514

375

308

626

6399

7:9

3015

422

4II

389

486

509

511,

046

1940

161

236

1240

949

853

754

1,08

919

5035

762

631

1,01

480

11,

422

142

2,36

519

5652

485

543

1,42

21,

168

1,94

319

43,

305

Sour

ces:

Col

. I: S

chul

tz (

21)

Tab

le 5

, CoL

(4)

div

ided

by

CoL

(1)

.C

oL 2

: Tab

le I

I ab

ove,

CoL

(3)

.C

oL 3

: Col

(2)

tim

esC

ol. 5

: Sch

ultz

(21

) T

able

6. C

ol. (

4) d

ivid

ed b

y C

oL (

1).

CoL

6: T

able

11

abov

e, C

ol. (

5).

Col

. 7: C

o!. (

6) ti

nes

10.


As for the third component, school-related costs Incurredby individuals, Schultz makes the assumption that these are5 per cent of foregone earnings for high school students and10 per cent of foregone earnings for college students.

By converting the school costs incurred by society to a percapita basis, we can compute the total resource costs of oneyear of high school and one year of college by summing thethree component costs. These are shown separately andtotalled in Table 12.

Schultz does not give us private resource costs, nor doeshe compute a rate of return. Hansen (14), however, fills thisgap, at least for the 1950 (school year 1949-50) data. He listsaverage tuition and fee charges as $245 per student in thatyear.

While Hansen is content to use Schultz's estimates ofschool costs and incidental school-related costs, he computeshis own estimates of foregone earnings. Using the same basicdata as Miller (17) and Houthakker (15), Hansen also com-putes "age-income profiles" for each level of education.(These "profiles" are shown in Table 1; relevant portions arereproduced here as Table 13.) Hansen states that "opportun-ity costs were taken directly from the age-income profiles ofthe alternative level of schooling being used in the calcula-tions. For example, at age eighteen the opportunity costs forthe person undertaking four years of college is the incomethat the high school graduate would obtain from ages eight-

Table 13Avta Act Incoms BY AGE AND YEARS or Sa1001,

COMPLETED, MAUS, UNITED STATES, 1949

AgeYears of School Cotnpleted

8 12

14-15 $ 40616-17 ...... 53418-19 1,069 $ 95S20-21 1,535 1,744

Source: Table 1.

( 24j

The Rate of Return on Investment in Education

een to twenty-one." (14, p. 130). Thus, the total opportunitycost of four years of college, as figures in Table 13 show,would be $5,398 (two years at $955 and two at $1,744), or anaverage annual cost of $1,349. Similarly, four years of highschool would involve opportunity costs of $1,880 (two yearsat $406 and two at $534), or $470 per year.

We can now combine the estimates thus far derived toshow total and private resource costs per student for 1949.This is done in Table 14. The figures shown here are slightlyless than those shown for 1950 in Table 12,

THE RATE OF RETURN

Using the cost data shown In Table 14 and the income datashown in Table 1, which he adjusts for mortality rates, Han-sen computes average and marginal rates of return on totaland private resource costs. For the latter he computes therates of return both before and after federal income tax. Hisresults are shown in Tables 1S, 16, and 17.

The diagonal elements in Tables 15-17 represent marginalreturns, while the off-diagonal elements are average returns.Hence we see that the profitability of investment in educationdepends upon one's time horizon. Consider Table 16, for ex-ample. If an individual has just completed grade one and isconsidering going on through two years of college (grade14), his average rate of return, before tax, would be 18.1 percent. If, however, he has just completed one year of collegeand is considering a second, he can expect a marginal rate ofreturn of only 6.2 per cent. The latter rate of return is con-ceivably less than could be received by investing in alternativeassets; the former is most likely higher. The time horizon Isalso important when considering the returns on total resourcecosts. Hence we see that public policy decisions based on thistype of analysis will differ, depending on whether averageor marginal rates are deemed to be relevant.

As was to be expected, the private rates, before tax, aregreater than the rates on total resource costs in all cases. This

( 25 1

Tab

le 1

4T

om A

ND

Pya

vAzz

Rno

nacz

Cos

ts ?

Ea

Srer

omvr

a Sa

lons

. Lev

y .,

1949

(D

otzA

vs)

Scho

olSc

hool

Cos

ts

Tar

ns. R

zsoc

acz

Cos

ts

Ear

ning

sIn

cide

ntal

Fore

gone

Cos

tsT

otal

Tui

tion

and

Fees

PRIV

AT

E R

ESO

UR

CE

CO

STS

Ear

ning

sIn

cide

ntal

.

Fore

gone

Cos

tsT

otal

Lev

elfg

,(1

)(2

)(3

)(4

)(5

)(6

)(7

)(8

)

te,

Ele

men

tary

201

201

3147

0H

igh

Scho

ol35

747

085

831

501

Col

lege

801

1,34

914

22.

292

245

1.34

914

21,

736

Sour

ce: C

ol. 1

: Sch

ultz

(21

), T

able

5, C

oL 4

div

ided

by

CoL

1.

CoL

2: D

ata

com

pute

d fr

om T

able

13.

Col

. 3: D

ata

com

pute

d fr

om T

able

12,

Inc

iden

tal c

osts

, 195

0.C

oL 5

: Han

sen

(14)

.C

oL 6

: Dat

a co

mpu

ted

from

Tab

le 1

3.C

oL 7

: Dat

a co

mpu

ted

from

Tab

le 1

2, I

ncid

enta

l cos

ts, 1

950.

Tab

le 1

5Ir

rrE

arem

. Rim

s O

F R

zru

7o T

arA

t. R

nouR

ce I

mE

srm

orr

SCH

OO

LIN

G. U

NH

-FD

ST

AT

ES,

MA

LE

s. 1

949

To:

From

:(I

)(2

)(3

)(4

)(5

)(6

)(7

)6

812

1416

1820

13

79

1113

15A

geG

rade

(1)

72

8.9

__

__

(2)

II6

12.0

14.5

____

__

(3)

138

15.0

18.5

292

(4)

1510

13.7

15.9

163

9.5

(5)

1712

134

15.4

153

11.4

137

_(6

)19

1411

.312

111

182

825.

4_

(7)

2116

121

12.7

12.1

10.5

10.9

102

154

All

rate

-of-

retu

rn f

igur

es a

re s

ubje

ct to

som

e er

ror,

sin

ce th

e es

timat

ion

to o

ne d

ecim

al p

lace

had

to b

e m

ade

byin

terp

olat

ion

betw

een

who

le p

erce

ntag

e fi

gure

s.So

urce

: Han

sen

(14)

, Tab

le 3

, 134

-

ft, a

Tab

le 1

6IN

TE

RN

AL

, Rat

s O

F R

ET

UR

N T

O P

RIV

AT

E R

imoO

RC

E I

NV

IMT

ME

NT

IN

SC

HO

OL

ING

BE

FOR

E T

AX

, UN

ITE

D S

TA

TE

S, M

AL

ES,

194

9'

From

:(1

)(2

)(3

)(4

)(5

)(6

)(7

)6

812

1416

1820

13

79

1113

15T

o:A

geG

rade

(1)

72

(2)

116

(3)

138

(4)

1510

283

34.6

25.9

(5)

1712

25.6

29.4

233

153

186

(6)

1914

18.1

183

14.8

10.4

9S62

(7)

2116

182

182

162

12.9

13.0

116

18.7

All

rare

-of

- re

turn

fig

ures

are

sub

ject

to s

ome

erro

r, s

ince

the

estim

atio

n to

one

dec

imal

pla

ce h

ad to

be

mad

e by

inte

rpol

atio

n be

twee

n w

hole

per

cent

age

figu

res.

+ T

his

indi

cate

s an

infi

nite

rat

e-of

-ret

urn,

giv

en th

e as

sum

ptio

n th

at e

duca

tion

is c

ostle

s to

the

indi

vidu

al to

the

com

-pl

etio

n of

eig

hth

grad

e.S

ourc

e:H

anse

n (1

4), T

able

4, 1

36.

Tab

le 1

7IN

TE

RN

AL

RA

TFS

OF

RE

TU

RN

TO

PR

IVA

TE

Itr

sout

ect I

sivz

stam

it r

&m

ow.=

Air

= T

Ax,

UN

rrm

ST

AT

ES.

MA

LIN

. 194

9'

.....-

..."-

-,T

o:Fr

om:

(1)

(2)

(3)

(4)

(5)

(6)

(7)

68

1214

16I8

201

37

911

13IS

Age

Gra

de

(1)

72

+(2

)11

6÷

4-(3

)13

8+

4-+

_ _

--(4

)15

1027

.933

.024

212

3

(5)

1712

252

282

22.2

145

17S

_(6

)19

1417

217

513

.79.

485

SI_

(7)

2116

172

173

14.4

115

11.4

10.1

16.7

All

rare

-of

-re

turn

fig

ures

are

sub

iect

to s

ome

erro

r, s

ince

the

estim

atio

n to

one

dec

imal

pla

ce h

ad to

be

mad

e by

inte

rpol

atio

n be

twee

n w

hole

per

cent

age

figu

res.

+ T

his

indi

cate

s an

infi

nite

rat

e-of

-ret

urn,

giv

en th

e an

umpt

ion

that

edu

catio

n is

cos

tless

to th

e in

divi

dual

to th

e co

m-

plet

ion

of e

ight

h gr

ade.

Sour

ce: H

anse

n (1

4), T

able

5, 1

36.


follows from the fact that total resource costs exceed privateresource costs. When comparing private rates, after tax, withtotal rates, however, we find two exceptions to this generalrule. After the completion of one year of college, both themarginal rate of return on a second year and the average rateof return on three more years are less for the individual thanthey are for society. This suggests, according to Hansen,"that the student pays more than his own way in securingschooling at the college level. This might indicate the needfor a re-study of the assesssment of the costs of college againstthe individual, unless the possible underinvestment in collegetraining that would be produced is regarded as acceptable insome broader sense." (p. 137)

By way of conclusion, Hansen compares the "rate-of-re-turn" approach with the "additional life-time income" ap-proach, both of which were examined above, in order to seehow an individull would rank various levels of schoolingconsidered as investments. For simplicity, he assumes thedecision maker to have just completed eight years of elemen-tary school and to be making one, presumably irrevocable, de-cision. As can be seen by looking at Table 18, the "additionallife -time income" approach suggests that any and all amountsof schooling are worthwhile; likewise, discounting at 3 and6 per cent gives the same impression. When discounting at 8or 10 per cent, however, two years of college are less profit-able than merely finishing high school, and the after tax re-turn to two years of college, when discounted at 10 per cent,is even less than the return to two years of high school. Withregard to the rate of return method, fouryears of high schoolappear to offer the highest average rate of return when onediscounts back to age 14. Of course, it is the marginal rateof return rather than the average rate of return upon whicha person bases his decision to continue his education.

' UNDERI NVESTM ENT IN COLLEGE EDUCATION"

The evidence which has been cited clearly indicates thatthe rate of return on a college education is in excess of 10 per

{ 30

Tab

le 1

8A

LT

ER

NA

TIV

E M

ET

HO

DS

OF

CO

MPA

RD

IG V

AL

VE

OF

PRIV

AT

E E

CO

NO

MIC

RE

rVR

NS

TO

Im

Esr

aftm

cr

IN S

CH

OO

LIN

GS,

AS

VIE

WE

D A

T A

CE

FO

UIT

TE

M, U

NT

IED

ST

AT

ES,

MA

LE

S, 1

949

Scho

olin

g fr

om C

orn-

pler

ion

of G

rade

8to

com

plet

ion

of

Add

ition

alL

ifet

ane

Inco

me

(1)

3 Pe

r ce

nt(2

)

Prits

wr

VA

LU

E o

r A

DD

ITIO

NA

L I

NC

OM

E A

r

6 Pc

r ce

nt8

Per

cent

10 P

er c

ent

(3)

(4)

(S)

Inte

rnal

Rar

e of

Ret

urn

(Per

cen

t)(6

)

.--,

Bef

ore

Tax

,...,

S 16

,802

$ 7,

756

$ 2,

301

S1,1

90$

545

12.7

- 2

year

s hi

gh s

choo

l __

_ __

-__

Ym

a4

year

s hi

gh s

choo

l46

,083

18,1

%6,

488

3,60

11,

949

153

2 ye

ars

colle

ge66

,763

23,8

007,

352

3,21

599

610

.4

4 ye

ars

colle

ge14

1,46

849

,429

17,2

528,

722

4,13

512

.9__

___

Aft

er T

ax

2 ye

ars

high

sch

ool

____

___

___

$ 14

,143

$ 5,

081

$ 1,

956

S 99

6$

436

123

4 ye

ars

high

sch

ool__

____

___

_ 38

,287

13,5

805,

362

2,92

91,

547

14.5

2 ye

ars

colle

ge--

-___

___

52,4

8517

,000

5,36

42,

084

336

9.4

4 ye

ars

colle

ge10

9,99

336

,575

12,8

246,

170

2,61

111

5

Sour

ce: H

anse

n (1

4), T

able

6,1

38.


cent. Furthermore, if we assume that the rate of return onalternative investments is approximately 5 per cent, then thesefigures clearly imply that there is underinvestment in collegeeducation. Becker (2) questions these percentages, however,stating: "Even 9 per cent is probably too high an estimate ofthe return to all college graduates since it refers only to urbanmale whites. The rate of return to nonwhites seems to beabout two percentage points lower than this."

Becker further discusses the average rate of return to busi-ness capital as dependent on the rates of return to ". thecorporate and unincorporated sectors and on the relative im-portance of each sector," Using these measurements he con-cludes ". . the average return to all business capital wouldbe 8 per cent." Becker's figures, then, would give us a fatdifferent picture of the rate of return on a college education.

1 321

The Contribution of Educationto Economic Growth

Another method of assessing the economic value of educa-tion, one which carries with it the implication of some rateof return, is that of measuring the contribution which edu-cation has made to economic growth. In this section we willconsider the estimates of this contribution made by Denison(9) and Schultz (21).

DENISON'S ESTIMATES

Using mean income data found in the work of Houthakker(15), Denison presents income differentials by level of educa-tion for males of the same age (see Table 19). Denison nextmakes the assumption that three-fifths of each of the differen-tials shown In column 1 of Table 19 are due to differences ineducation. He then derives new differentials which reflectonly this difference. These are shown in column 2 of Table19 (col. 2 = 3/5 [col. 1 100] + 100).

Denison's assumption made it possible for him to calculatethe effects of increased education on past growth. For eachyear for which he could derive a distribution of individualsby number of years of school completed, he calculated whatthe average earnings of males over 25 would have been if theearnings at each educational level were a constant fraction(column 2, Table 19) of actual 1949 earnings of eighth gradegraduates. "The differences from period to period of earn-ings so computed can be used to isolate the effect of changesin the length of schooling, measured in years, on average in-come. An adjustment is then possible to take account ofchanges in the number of days of school attendance duringthe year." (p. 70)

[33]


Table 19

MEAN INVME DIFFERENTIALS BY LIASLOP SCHOOLING COMPLETED

Years of SchoolCompleted

(1)Meali Income as

% of Mean Incomeof Eighth Grade

Graduates

(2)Mean Income DifferentialUsed to Represent Effect

of Education (% ofIncome of EighthGrade Graduates)

None ......... ....

Elementary School1 to 4 yearsS to 7 years

8 yearsHigh School

1 to 3 years 115

4 years 140

College1 to 3 years ...... 165

4 years or more 235

........

SO 70

65

80

100

79

88

I00

109

124

139

181

Source: Denison (9), Table 8, 68.

His results are shown in Table 20. (It should be noted thatthis adjustment is based on the assumption that a doubling ofthe number of days of school attended per year while hold-

Table 20LABOR OUTPUT PER MAN BASED ON TOTAL. DAYS OF EDUCATION

PeriodAnnual Rate of Change

Per Cent Change (Per Cent)

1910 to 1920 4.9 0.481920 to 1930 6.9 0.671930 to 1940 8,8 0.851940 to 1950 10.4 1.001950 to 1960 10.3 0.991910 to 1930 12.1 0.571930 to 1960 32.6 0.941910 to 1960 48.6 0.79

Source: Denison (9) Table 9, p. 72, Cols. S and 6.

( 34 1

The Contribution of Education to Economic Growth

ing the number of years constant would have the same effecton output as a doubling of the number of years of schoolwhile holding the number of days per year constant.)

The meaning of the results can be seen by considering theperiod 1950 to 1960. The first column of Table 20 tells usthat "if the labor force in 1950 had been as well educated asthat of 1960, it would have contributed 10.3 per cent more toproduction than it actually did. Since labor represented about7S per cent of the national income at that time, the nationalincome would have been larger by 7.7 per cent." (9, p. 72)

The second column of Table 20 shows the average annualrates of change which are implied by the total period changesin the first column, Thus, if we consider the period 1930 to1960, we see that output per laborer due to education in-creased at an average annual rate of ,94 per cent. Nationalincome in this period was increasing at approximately 3 percent per year (see 9 and 21). Hence, again assuming labor'sshare to be 75 per cent we find that increased education con-tributed .75 X .94 = .705 points to the growth rate of out-put. In other words, 23.5 per cent of the growth experiencedfrom 1930 to 1960 was due to increased education of thelabor force.

In his analysis Denison focuses attention on the period from1929 to 1957. He concludes that the contribution of educa-tion to economic growth during these years was 23 per cent."When related to the growth of national product per personemployed, the contribution of additional education appearsstill more impressive. My final estimate is that education con-tributed 42 per cent of the 1.60 percentage point growthrate in product per person employed." (p. 73)

The importance of Denison's assumption that three-fifthsof the income differential was due to differences in educationcannot be stressed too strongly. The effect of alternative as-sumptions can he approximated by multiplyin3 the resultsin. Table 20 by the ratio of the alternative percentage to 60per cent. Thus, if we assumed that 75 per cent of the differ-ential was due to differences in education, we would credit

I 35 )


29.3 per cent of the growth in output to education. Similarly,substitution of 50 per cent would credit 19.6 per cent of totalgrowth to education.

SC I !ULU'S ESTI NI AT ES

According to Schultz (21) the stock of education carriedby the labor force was equal to 180 billion dollars in 1929and 535 billion dollars in 1957, both measured in 1956 prices.This represents an Increase of 355 billion dollars. Meanwhile,the real income of the United States rose from 150 to 302billion dollars in 1956 prices during this period. If we againMime labor's share to be 75 per cent, then its contribution tooutput increased from 112.5 to 226.5 billion dollars over theperiod. If earnings per person had been held constant, laborwould have earned only 155.5 billion dollars in 1957. Hence,they earned 71 billion dollars more than they would have ifearnings per laborer had been held constant at the 1929 level.Schultz now asks the question: "How much of this 71 billiondollars is attributable to more education?"

Because the labor force Increased by 38 per cent between1929 and 1957, Schultz adds 69 billion dollars to the 1929stock of education (38 per cent of 180 billion dollars) in orderto keep the per laborer stock of education constant in thesetwo years. Hence, the 355 billion dollars increase in the stockof education carried by the labor force consists of two parts.The first part, 69 billion dollars, is due to growth in the laborforce. The second part, 286 billion dollars, is due to an In-crease in the stock of education per laborer.

The extent to which these two increases contributed tothe growth in national income clearly depends upon the rateof return earned by this investment. Schultz uses three esti-mates, as shown in Table 21. The 9 per cent rate he takes fromBecker (2). The I 1 per cent is his own estimate of the returnto college education in 1958, and the 17.3 per cent is aweighted average of his estimates for the returns to elemen-tary, high school, and college education (35, 10, and 11 percent, respectively). His results are summarized in Table 21.

[ 36

Tab

le 2

1E

srim

Azz

s or

TH

E C

omni

onyn

ow n

t TH

E L

amm

Foa

m T

o E

atim

ios

Aim

zo

IsT

Azz

oNA

L I

mcw

.te, B

Erw

zor

1929

AN

D 1

956,

or

Tue

Uur

nto

STA

=

Rat

e of

Ret

urn

Stoc

k of

edu

catio

n ad

ded

(in

billi

ons

of d

olla

rs)

Part

One

Part

Tw

o

Inco

me

attr

ibut

able

toth

is a

dditi

onal

edu

catio

nbi

llion

s of

dol

lars

)Pa

rt O

nePa

rt T

wo

Tot

al

Prop

ortio

n of

the

71 B

illio

n D

olla

rsin

crea

se in

ear

n-in

gs o

f la

bor

coL

Sit

100

71

(1)

(2)

(3)

(4)

(5)

(6)

(7)

969

286

6225

.731

936

11

6928

674

31S

39.1

4417

369

286

11.9

49S

614

70

Sour

ce: S

chul

tz (

21),

Tab

le 1

8 81

.


Thus, the increase in education per member of the laborforce "explains" between 36 and 70 per cent of the otherwiseunexplained increase in earnings per laborer.

Comparing these results with Denison's, we first note thatthe n billion dollars increase in labor's earnings representsroughly 46 per cent of the growth of national income. Apply-ing the 36 and 70 per cent figures to this, we see that the in-crease in education per laborer accounted for between 16.6and 32.2 per cent of the growth in income from 1929 to 1957.

l3Hi

VSummary

This monograph deals with the findings of economists con-cerning education as an investment. Data are presented onthe benefits accruing both to individuals and to the nation asa whole from investment in education. In summarizing ourfindings, we shall state generalizations pertaining to (1) theeffects of education on earnings, (2) the rate of return oninvestment in education, and (3) the contribution of edu-cation to economic growth.

EDUCATION AND EARNINGS

1. Data on annual income for males show that at all agelevels income increases as years of schooling increase. Forexample, in the category from 25 to 34 years, average annualincome in 1958 was $3,663 for elementary school graduates,$4,909 for high school graduates, and $7,152 fortollege grad-uates. income for persons in the years of peak earning power,the 45 to 54 age bracket, ranged from $4,337 for elementaryschool graduates to $6,295 for high school graduates and$12,269 for college graduates.

2. The relationship between income and educational at-tainment has persisted through the years, even though theamount of school attained by the population has increased.indeed, income differentials between elementary school grad-uates and high school graduates, and between high schooland college graduates have increased in recent years. In 1949,the average high school graduate had an income 34 per centgreater than the elementary school graduate, while the collegegraduate's income exceeded that of the high school graduateby 63 per cent. By 1958, the high school graduate's advantageover the elementary school graduate had increased to 48 per

39]


cent, and that of the college graduate over the high schoolgraduate to 65 per cent.

S. In terms of income received, persons with the most edu-cation benefit to the greatest extent from years of job experi-ence, The college graduate benefits most in this respect; hisearnings increase proportionately more than do those of ele-mentary and high school graduates as he gets older. For ex-ample, in 1958 college-trained individuals in the 4S to 54 agebracket had average incomes that were 72 per cent higherthan those of persons in the 25 to 34 age category. Compar-able differences for elementary school and high school grad-uates were 18 per cent and 28 per cent, respectively. Thusthe college graduate, who earns more than persons with lesseducation at any age level, is increasingly advantaged as hegrows older.

4. Total lifetime income increases as education increases.As of 1958, the average mate graduate of elementary schoolcould be expected to earn $169,976 during his lifetime, ascompared with $241,844 for the high school graduate and$419,871 for the college graduate.

5. As in the case of annual income, the greatest gains inlifetime income in recent years have accrued to persons withthe most education. For example, in 1946 high school gradu-ates could expect to earn 35 per cent more than elementaryschool graduates; by 1958, they could expect to earn 44 percent more. The same pattern, but in far more pronouncedform, exists between high school and college graduates. Thedifferential in favor of college graduates was 48 per cent in1946, but it had increased to 70 per cent by 1958.

6. Even when lifetime income is discountedthat is,equated to return on current investmentthe contribution ofadditional education to earnings is positive and significant.

THE RATE OF RETURN ON INVESTMENT IN EDUCATION

1. Education yields a high rate of return on investment

( 40 j

Summar

i.e., the monetary returns exceed the costs of education by aconsiderable margin. This is true from the point of view.ofsociety as a whole as well as that of the individual who In-vests in education for himself. The benefits to society areimpressive; for example, in 1949 an investment in educationthat would have permitted male first graders to complete highschool would have produced a 13.6 per cent return, on theaverage. An investment through four years of college wouldhave returned 12.1 per cent on the total investment. Evenmore striking are the economic returns to individuals. Forexample, the male first grader in 1949 could expect a 25.6per cent return on the private funds required to see himthrough high school, and a return of 18.2 per cent on thefunds required for him to graduate from college.

2. Research on additional lifetime income resulting fromprivate investment in more education shows that the rate ofreturn remains high at all educational levels. For example,for males In 1949 the rate of return to funds used to educatethe eighth grader through four years of high school was 15.3per cent; the return to an investment in four years of collegewas 12.9 per cent.

3. It is quite possible that society (as well as individuals) ismaking an "underinvestment" in college education. This con-tention rests upon the fact that the rate of return on a collegeeducation is clearly in excess of 10 per cent, while the rateof return on alternative investments is considerably lower(approximately five per cent).

THE CONTRIBUTION OF EDUCATION TO ECONOMIC GROWTH

I. Studies have concluded that increases in years of school-ing completed by the labor force have contributed signifi-cantly to the economic growth of the nation. Denison, study-ing data for the period from 1929 to 1957, found that 23 percent of the economic growth of the country during theseyears was attributable to the increased education of the laborforce. Schultz, who studied increases in income in the United

[41)


States during the same years, concluded that the increase ineducation per laborer accounted for between 16.6 and 32.2per cent of the growth in income during this period.

2. These studies reinforce the conclusion reached In previ-ously mentioned research dealing with the rate of return onInvestment In educationthat is, the social returns on educa-tional investment are very profitable, as are private returnson Individual investment. Education is, then, not merely anexcellent investment for the individual; it also yields heavyreturns to the society as a whole,

42 1

BIBLIOGRAPHY

1. Abramovitz, Moses; "Economic Growth in the UnitedStates," American Economic Review, Vol. S2 (Sept.,1962), pp. 762-782.

2. Becker, Gary S.; "Underinvestment in College Educa-don?", Americo; Economic Review, Vol. 50 (May,1960), pp. 346-354.

3. ; "Investment in Human Capital: A TheoreticalAnalysis," Journal of Political Economy (Supplement),Vol. 70 (Oct., 1962), pp. 9-49.

4. Benson, Charles S. (ed.); Perspectives on the Economicsof Education, Boston: Houghton Mifflin Company, 1963.

S. Bowman, Mary Jean; "Human Capital: Concepts andMeasures," In The Economics of Higher Education, ed.Selma J. Mushkin. Washington; U.S Department ofHealth, Education and Welfare, Office of Education,1962, pp. 69-92.

6. ; "Social Returns to Education," InternationalSocial Science Journal, Vol. 14 (Dec., 1962), pp. 647-659.

7. "Converging Concerns of Economists andEducators," Comparative Education Review, Vol. 6(Oct., 1962), pp. 111-119.

8. Debeauvais, M.; "The Concept of Human Capital," In-ternational Social Science Journal, Vol. 4(Dec., 1962),pp. 660-675.

9. Denison, Edward F.; The Sources of Economic Growthin the finked States and the Alternatives Before Us.,New York: Committee for Economic Development,1962.

10. "Education, Economic Growth and Gaps inInformation," Journal of Political Economy (Supple-ment), Vol. 70 (Oct., 1962), pp. 124-128.

11. Ecliaus, Richard S.; "Education and Economic Growth,"in The Economics of Higher Education, ed. Selma J.

43 )


Mushkin. Washington: U.S. Department of Health, Edu-cation and Welfare, Office of Education, 1962, pp. 102-128.

12. ; "Economic Criteria for Education and Train-ing," Review of Economics and Statistics, Vol. 46 (May,1964), pp. 181-190.

13. Glick, P. C. and B. P. Miller: "Educational Level andPotential Income," American Sociological Review, Vol.21 (June, 1956), pp. 307 -312,

14. Hansen, W. Lee; "Total and Private Rates of Return toInvestment in Schooling," Journal of Political hcononty,Vol. 71 (April, 1963), pp. 128-140.

15. Houthakker, H. S.; "Education and Income," .Review ofEconomics and Statistics, Vol. 41 (Feb., 1941), pp. 24-28.

16. Hunt, Shane J.; "Income Determinants for College Grad-uates and the Return to Educational Investment," YaleEconomic Essays, Vol. 3 (Fall, 1963), pp. 305-357.

17. Miller, FI. P.; "Annual and Lifetime Income in Relationto Education: 1939-1959," American Economic Review,Vol. SO (Dec., 1960), pp. 962-986.

18. ; "Income and Education: Does Education PayOff ?", in The Economics of Higher Education, ed. SelmaJ. Mushkin. Washington: U.S. Department of Health,Education and Welfare, Office of Education, 1962, pp.129-146.

19. Mincer, Jacob; "On -The -Job Training: Costs, Returns,and Some Implications," Journal of Political Economy(Supplement), Vol. 70 (Oct., 1962), pp. 50-79.

20. Renshaw, Edward F.; "Estimating the Returns to Edu-cation," Review of Economics and Statistics, Vol. 42(Aug., 1960), pp. 318-324.

21. Schultz, T. W.; "Capital Formation by Education," Jour-nal of Political Economy, Vol. 68 (Dec., 1960), pp. 571-583.

22. ; "Education and Economic Growth," in Social

[44)

Bibliography

Forces influencing American Education, ed. Nelson B.Henry. Chicago: University of Chicago Press, 1961, pp.46-88.

23. ; "Investment in Human Capital," AmericanEconomic Review, Vol. 51, (March, 1961), pp. 147.

24. ..... "Investment in Human Capital: Reply," Amer-ican Economic Review, Vol. 51, (Dec., 1961), pp. 1035-1039.

25. ; "Reflections on Investment in Man." Journalof Political Economy (Supplement), Vol. 70 (Oct.,1962), pp. 1-8.

26. ; "Rise in the Stock of Capital Represented byEducation in the United States, 1900-57," in The Eco-nomics of Higher Education, ed. Selma J. Mushkin.Washington: U.S. Department of Health, Education andWelfare, Office of,Education, 1962, pp. 93-101.

27. ; The Economic Value of Education, NewYork: Columbia University Press, 1963.

28. Shaffer, Harry; "Investment In Human Capital: Com-ment," American Economic Review, Vol. 51 (Dec.,1961), pp. 1026-1034.

29. Vaizey, John; The Economics of Education, New York:Free Press of Glencoe, Inc., 1962.

30. Walsh, J. R.; "Capital Concept Applied to Man," Quar-terly Journal of Economics, Vol. 49 (Feb., 1935), pp.255-285.

31. Weisbrod, Burton A.; "The Valuation of Human Capi-tal," Journal of Political Economy, Vol. 69 (Oct., 1961),pp. 425.436.

32. ; "Education and Investment in Human Capi-tal," Journal of Political Economy (Supplement), Vol.70 (Oct., 1962), pp. 106-123.

45

document resume ed 089 396 24 81 005 983document resume 24 81 005 983 innes, jon t.; and others the...

Documents