other determinants of professional income
TRANSCRIPT
This PDF is a selection from an out-of-print volume from the NationalBureau of Economic Research
Volume Title: Income from Independent Professional Practice
Volume Author/Editor: Friedman, Milton and Simon Kuznets
Volume Publisher: NBER
Volume ISBN: 0-87014-044-2
Volume URL: http://www.nber.org/books/frie54-1
Publication Date: 1954
Chapter Title: Other Determinants of Professional Income
Chapter Author: Milton Friedman, Simon Kuznets
Chapter URL: http://www.nber.org/chapters/c2329
Chapter pages in book: (p. 235 - 299)
DETERMINANTS OF PROFESSIONAL INCOME 235
physicians as dentists in active practice in the United States. Inthe West South Central region the corresponding ratio was 3.18and in the West North Central, i.78.'° In the West South Centralregion physicians are numerous relatively to dentists and theirincomes are low relatively to dentists'; in the West North Centralregion the relation is reversed. These figures are merely suggestiveand do not conclusively establish that the observed difference inrelative incomes reflects this difference in the relative number ofpractitioners; indeed, one region, the East South Central, has aneven higher ratio of physicians to dentists, 3.27, than the WestSouth Central region; and one region, the Pacific, has an evenlower ratio, i.6i, than the West North Central.
CHAPTER 6
Other Determinants ofProfessional Income
1 TRAINING AND ABILITY
THE KIND OF TRAINING individuals get and the ability they pos-sess play a large role in determining their professional com-petence, connections, and opportunities; and, through these,their incomes. Unfortunately, data for measuring the influenceof these important factors are almost nonexistent. The onlyinformation available on the influence of training and abilityis from two fragmentary studies of lawyers, one for New YorkCounty, the other for Wisconsin. The New York County study
10 See Appendix to Chapter 4, Section 3b, for a more detailed discussion of therelation between the incomes of physidans and dentists and the number ofpractitioners, and for the source of these figures.
236 PROFESSIONAL INCOMEpresents data, summarized in Table 30, on the incomes ofcollege graduates and nongraduates and of graduates of full-and part-time law schools. The classifications are of course in-terrelated: individuals who graduate from college are morelikely to attend a full-time law school than those who do not.
The median income of college graduates is almost 50 percent larger than that of noncollege graduates; similarly, themedian income of graduates of full-time law schools is about
TABLE 30
Median Income, by Prelegal Trainingand by Type of Law School Attended
New York County Lawyers, and 1928—1932
NO. OF LAWYERS
REPORTING INCOME FOR MEDIAN INCOME
1933 1928—32 1933 1928—32
(dollars)
Prele gal trainingCollege graduate 1,742 1,317 3,570 5,220Not college graduate 1,026 795 2,480 3,570
Type of law school attendedFull-time 1,794 1,402 5,140Part-time i,oii 786 2,580
Suruey of the Legal Profession in New York County (New York County LawyersAssociation, 1936), p. 23.
50 per cent larger than that of graduates of part-time lawschools. These results presumably indicate that more extensivetraining yields higher incomes. However, the results are notunambiguous. On the one hand, the increasing tendency forindividuals to go to college means that college graduates areconcentrated in younger age groups than other lawyers. Sinceincome tends to increase for a time with age (see Sec. 2), themedians in Table 30 probably understate the difference be-tween the incomes of college graduates and others in practicethe same number of years. On the other hand, both collegegraduates and graduates of full-time law schools doubtlesscome from wealthier families and have more valuable profes-sional connections than other lawyers. The latter frequently
DETERMINANTS OF PROFESSIONAL INCOME 237
cannot afford to go to college and often attend part-timeschools while earning their living. Hence, lawyers who re-ceived the better training would probably have earned higherincomes than their fellow lawyers even if they had receivedthe same training as the latter.
The Wisconsin data relate 1932 incomes to the standing ofindividuals in their law school classes (Table si). In general,
TABLE 31
Professional Income, by Grade in Law School
Graduates of University of Wisconsin Law School
OF LAWYERS ATrAININC SPECIFIFDGRADE IN LAW SCHOOL WHOSE 1932 INCOME WAS
GRADE IN NO. OF LAWYERS $o— $2,ooo— OverLAW SCHOOL REPORTING INCOME Loss 2,000 4,000 $4,000
A 67 .0 29.9 32.8B '4,3 .7 34.9 32.2 32.2C .7 41.2 32.4 25.7D 113 41.6 29.2 28.3E. 137 5.1 50.4 29.2 15.3
L. K. Garrison, 'A Survey of the Wisconsin Bar', Wisconsin Law Review, Feb.1935, pp. 161—3. Data are for the graduating classes of University of WisconsinLaw School, 1904—31, excluding 1917—20.
those who received high grades in law school had better suc-cess in the practice of law than those who received low grades,although the correlation is far from close. The looseness of therelation betweenlaw school grades and income may mean thatthere is equally little connection between ability and income;probably, however, it means that law school grades a poorindex of the type of ability that spells success in the p1 actice oflaw. High grades may measure application rather than ability,and individuals may 'loaf' in law school but work h;i.rd oncethey enter practice. In addition, the type of ability toget high grades may not be the same as the type of abilityneeded to make money as a lawyer.
2 NUMBER OF YEARS IN PRACTICE
The early career of an independent professional man is char-acterized by low earnings. Some time must elapse before his
238 PROFESSIONAL INCOMEavailability and competence become known—before he canattract 'practice'. To get a client is as a rule more difficult thanto keep him. Once started, his practice tends to be cumulative.A satisfied client will return again and again and will recom-mend the practitioner to his friends or business associates. Inthe course of time the professional man becomes known, gainsa reputation, acquires experience—that is to say, 'builds up apractice', and therefore is able to charge higher fees and tokeep more fully occupied. If his reputation is good, his prac-tice and income will continue to grow. But a limit to growthis set by the amount of business he can handle, the size of themarket he serves, and the competition of fellow practitioners.This limit may come early or late in a man's professional ca-reer. But sooner or later he will have to struggle to retain hisclients and to replace those he is unable to retain. And even-tually, he will pass the zenith of his career. Long before hisphysical energies decline, he will find it increasingly difficult toretain his practice in the face of competition of younger men,trained in the latest methods and vigorously striving to makea place for themselves. Some of the more eminent, of course,will easily weather this competition and continue to increasetheir practice and income long after most other practitionersof their age are experiencing serious difficulty in maintainingtheir position. But others will find their problem intensifiedby poor health, while still others will voluntarily relinquishtheir practice in whole or in part as the urge for retirementovercomes the drive for professional advancement.
The data summarized in Table 32 and Charts 20 and 21document this picture of an initially rapid rise in average in-come, followed by a slower rise, relative stability, and ulti-mately, a decline. In medicine and dentistry, professions inwhich scientific advance has been rapid and in which physicalskill and dexterity are required, younger men are at an ad-vantage and the peak income is reached fairly early—betweenthe thirteenth and twentieth year of practice. In law, the onlyone of the 'business' professions for which data are available,experience and contacts are more important and physical fit-
CHART 20
Arithmetic Mean Income of Physicians and Dentists
by Years in Practice
Arithmetic mean incoms(dollars)6500
6003
I I I I I I I I I I I
5 tO 15 20 25 30 35 40 45Years in practice
50 55 60
The midpoint for California dentists '40 and over' class has been arbitrarily set at 5o; forCalifornia physicians and over', at
ness secondary. In consequence, peak income is reached muchlater. Indeed, according to the data for New York County, theoldest lawyers have the highest median income. However, thelast class—in practice 35 years and over—is fairly broad and ifsubdivided might well reveal a decline. Wisconsin lawyersseem to attain peak incomes between the twentieth and forti-eth year of practice.
DETERMINANTS OF PROFESSIONAL INCOME 289
... S
• S
..•5
.
5500
5000
4500
4000
3500
3000
2500
2000
1500
1000
500
0
S
S
CaliFornia dentists, 1933
CaRlornia physicians, 1933
Dentists in 20 states, 1929
Physici.ans in U. S., 1936 (1937 sample)
0
J0 0
TA
BLE
32
Ave
rage
Inc
ome,
by
Yea
rs in
Pra
ctic
e
Phys
icia
ns, D
entis
ts, a
nd L
awye
rs: S
elec
ted
Stud
ies
PHY
SIC
IAN
S &
DK
NT
ISrS
INC
AL
IVO
PNIA
,1 1
953
INC
OM
&
PHY
SIC
IAN
S IN
L,1
PH
YS
ICIA
NS
DE
NT
IST
SD
EN
ThT
S IN
20
STA
T2S
,IN
CO
ME
iIN
CO
ME
Ari
th.
Aut
h.E
quiv
. no.
Ari
th.
Yea
rs in
Ari
th.m
ean
Yea
rs in
Med
ian
mea
nM
edia
nm
ean
Yea
r of
of y
ears
Med
ian
mea
n
prac
tice
inco
me
prac
tice
inco
me
inco
me
inco
me
inco
me
grad
uatio
nin
pra
ctic
ein
com
ein
com
e
(dol
lars
)(d
olla
rs)
(dol
lars
)(d
olla
rs)
Und
er s
1,78
8U
nder
*6o
o61
626
1929
—25
Und
er 6
2,69
53,
025
3—7
3,18
22
1,10
01,
175
j'
1924
—20
6—10
4,04
94,
742
8—i*
4,71
83
1,50
01,
771
1,50
03,
642
1919
-15
11—
155,
047
3,97
6
13—
175,
326
42,
000
2,06
81,
500
3,85
719
14-1
016
-20
4,98
*3,
995
18—
225.
179
5— 9
2,95
03,
267
2,40
02,
674
1909
-05
21—
254,
741
5,53
1
23—
274,
859
10-1
43,
600
4,46
52,
900
3,16
819
04-0
026
-30
4,26
13,
238
28—
3*3,
575,
15—
193,
600
4,50
83,
000
3,47
437
700
33—
374,
076
20—
243,
500
4,39
92,
500
2,96
218
94—
9036
—40
4,83
9z
38-4
23,
091
25—
292,
900
3,82
42,
500
2,82
841
—45
4,16
73,
940
43-4
72,
402
30—
343,
000
4,22
72,
200
2,48
018
8480
4650
5,12
53,
682
48—
521,
703
35—
392,
300
2,93
82,
100
2,20
0
53—
5774
740
-44
1,90
02,
838
—45
—49
1,40
02,
048
3,91
1Z
Ove
r49
1,15
01,
344
J
All
5,94
12.
700
3,57
22,
500
2,76
94,
094
3,01
1
No.
of
retu
rns
in s
ampl
e1,
409
2,73
71,
615
4.70
512
1
LA
WY
D.S
IN
NE
W Y
OR
E C
OU
NT
Y,'
2933
IN
CO
ME
Yea
r of
Equ
iv. n
o.ad
mis
sion
of y
ears
Med
ian
to b
arin
pra
ctic
ein
com
e(d
olla
rs)
Aft
er 1
929
Und
er1,
794
1924
—29
5—10
3,16
42g
18—
s311
—16
5,54
719
11—
1717
—53
6,67
419
00—
1024
—54
6,57
2B
efor
e 19
00O
ver
8,o6
8
AU
LA
WY
ER
S IN
WIS
CO
NSI
N,5
193
0 IN
CO
ME
AuT
h. M
EA
N I
NC
OM
E O
F L
AW
YE
RS
PRA
CT
ISIN
O I
NC
omm
uniti
es w
ith p
opul
atio
ns o
fY
ears
inO
ver
io,o
oo-
2,oo
o-U
nder
prac
tice
Milw
auke
e25
,000
610
,000
*,oo
o(d
olla
rs)
Und
er 5
1,77
*2,
224
1,98
31,
937
1,70
95—
93,
196
3,84
93,
764
3,18
42,
509
10—
195,
035
5,26
43,
223
20—
297,
846
6,95
05,
495
4,28
63,
422
30—
3410
,415
6,so
i5,
652
3,75
92,
512
35—
394,
415
4,29
02,
235
Ove
r 39
5,04
95,
558
3,71
62,
554
3,36
4
2,93
8
I-i
fTJ z z 0 IT,
No.
of
retu
rns
in s
ampl
e8i
*52
518
629
216
1
1 Se
e T
able
Cal
ifor
nia
Med
ical
-Eco
nom
ic S
urve
y,Fo
rmal
Rep
ort
on F
actu
al D
ata
(Cal
ifor
nia
Med
ical
Ass
ocia
tion,
pp. S
o, 8
8.'M
auri
ce L
even
, Pra
ctic
e of
Den
tistr
y an
d th
e In
com
esof
Den
tists
in
Tw
enty
Sta
tes:
192
9(U
nive
rsity
of C
hica
go P
ress
, 193
2), p
. 135
.'S
urve
y of
the
Leg
al P
rofe
ssio
n in
New
Yor
k C
ount
y, p
. 20.
5Gar
riso
n, 'A
Sur
vey
of th
e W
isco
nsin
Bar
', p.
156
.•
Exc
ludi
ng M
ilwau
kee.
co 0 z t-I z c-s 0
CHART 2t
Average Income of Lawyers by Years in Practice
New York County, median income in 1930Wisconsin, artthmetic mean income in 1930
Milwaukeeover 25,000 (excl. Milwaukee)
Communities of 10,000— 25,000Communities of 2,000— 10,000
— "'—'"—Communities under 2,000
Average net income(dollars)
11000
10000
55 60
The midpoint for Wisconsin 'over 40' years in practice has been arbitrarily set at forNew York 'over at 45.
Both the rise and decline seem larger for physicians thandentists. The factors responsible for this difference are pre-sumably the same as those which account for the greater vari-ability of medical than of dental incomes (see Ch. 4, Sec. 3).The most significant in the present connection is probably thegreater importance attached to medical services and the con-sequent willingness to pay more for the services of a physician
242 PROFESSIONAL INCOME
9000
8000
7000
6000
5000
4000
3000
2000
1000
0
..—...-..-."....%.
••_...__..._.. ..._..._•••e
I I I I I I I I
0 5 10 IS 20 25 30 35 40 45 50
Years in practice
DETERMINANTS OF PROFESSIONAL INCOME 243
considered superior to another than for the services of a den-tist considered superior.
For physicians, information on years in practice from the1937 sample makes possible a more intensive analysis.1 Thepercentage distribution of physicians by years in practice inTable is bimodal. The first mode is in the 3—7 years-in-prac-tice class; the second, in the 23—27 years-in-practice class. Theclass containing the second mode includes individuals whograduated from medical school about igio. It will be recalledfrom Chapter 1 that the initiation shortly before this date of
1 The available data on years in practice are not unambiguous. The questionwas 'state the number of years you have practised medicine". No other instruc-tions were given. Consequently it is not clear whether the respondents excludedtime spent as interns or in postgraduate study, or other interruptions to prac-tice. However, this vagueness cannot be important; with rare exceptions, themaximum possible error is one or two years. For reasons of economy, only thedata on income in 1936 have been analyzed by number of years in practice.
The questionnaire sent to lawyers in 1937 contained a question on the numberof years in practice. but we have made no use of the answers. We have re-peatedly had occasion to question the validity of this sample; in addition,the tabulation of the answers would be exceedingly because of thenecessity of both adjusting for the size of community bias and weighting theindividual states in combining them. An additional difficulty is the question-naire used. It was in two parts: the upper half requested data about the legalenterprise, the lower half about the individual lawyer. Additional copies ofthe lower half were enclosed with questionnaires sent to firm members, whowere asked to distribute them among the other members of the firm. Almostall individual practitioners who replied returned both halves; and so did mostfirm members who returned the upper halves. In many cases, however, addi-tional copies of the lower half were not returned by the other firm members.There is therefore reason to suppose that the sample of completed 'lowerhalves' is even more biased than the entire sample.
A table of average income of lawyers by years in practice based on this 1937Department of Commerce sample, which did not seem to us worth tabulatingin that fashion, is presented in Economics of the Legal Profession (AmericanBar Association, 1938), p. 21. So far as we can gather, the fIgures presented arebased solely on the lower half of the schedules and are not corrected for thesize of community bias or the nonrandomness of the sample by states. As a con-sequence the average 1936 income of all lawyers computed from this table is$3,446, whereas the average we obtain for 1936 from the same sample is $5,202(Table io). The table shows a rising income through 28—37 years in practice,and a decline thereafter.
10
TA
BLE
33
Perc
enta
ge D
istr
ibut
ion
by Y
ears
in P
ract
ice;
Ari
thm
etic
Mea
n In
com
e an
d N
umbe
r of
Per
sons
Cov
ered
,by
Yea
rs in
Pra
ctic
e an
d T
ype
of P
ract
ice
Phys
icia
ns, 1
936
AR
ITH
ME
TIC
ME
AN
INC
OM
EA
UG
ener
alPa
rtia
l%
DIm
InU
TIO
N'
phys
icia
nspr
actit
ione
rssp
ecia
lists
(dol
lars
)
YE
AR
S IN
PR
AC
TIC
E
Und
er3
3— '7
8—12
13—
17
18—
22
23—
27
28—
32
33—
3738
—42
43—
4748
—52
53—
57
AU
dns
ses'
6.7
1,78
81,
770
1,72
217
.33,
182
3,12
43,
085
13.8
4,71
83,
813
4,57
1
9.9
5,32
64,
071
4,90
88.
25,
179
3,50
04,
744
12.1
4,85
93,
261
5,03
211
.23,
575
2,60
53,
980
9.1
4,07
62,
538
4,28
26.
63,
091
2,25
72,
766
2.8
2,40
21,
967
2,23
6i.6
1,70
31,
169
2403
0.7
747
692
922
Com
plet
esp
ecia
lists
i,8g6
3,54
06,
133
7,31
2
7499
6,74
24,
558
7,15
5
10,7
49
NU
MB
ER
OF
PER
SON
S C
OV
ER
ED
1A
llG
ener
alPa
rtia
lC
ompl
ete
phys
icia
nspr
actit
ione
rssp
ecia
lists
spec
ialis
ts
8i37
2420
228
114
7617
268
4112
537
5210
534
4328
179
726s
46
167
71
o14
662
58
2610
030
12IT
I
3319
122
126
75
2
1,40
951
329
3
mat
e th
e pe
rcen
tage
dis
trib
utio
n of
phy
sici
ans
by y
ears
inpr
actic
e.Z
1lncludes
som
e re
turn
s fo
r w
hich
the
num
ber
of y
ears
in p
rac-
tice
was
unk
now
n.
100.
03,
941
2,91
73,
957
5,79
5
1Exd
udes
ret
urns
for
whi
ch th
e nu
mbe
r of
yea
rs in
pra
ctic
ew
as u
nkno
wn.
'Num
ber
of p
erso
ns c
over
ed b
y th
e re
turn
s us
ed b
efor
e w
eigh
t-in
g. T
hese
num
bers
, the
refo
re, c
anno
t be
used
dir
ectly
to e
sti-
DETERMINANTS OF PROFESSIONAL INCOME 245
an intensive drive for higher standards of medical educationcaused a sharp decrease in medical students and graduates.
Table and Chart 22 give the average 1936 incomes of allphysicians, general practitioners, partial specialists, and corn-
CHART 22
Arithmetic Mean Income of Physicians by Type of Practice
and Years in Practice, 1936
Arithmetic mean income.(dollars)1Z003
11000
10000
9000
8000
7000
6000
5000
4000
3000
2000
1000
0
plete specialists by years in practice. According to the averagesfor all physicians, it takes approximately seven years for physi-cians to attain an average income equal to theaverage for theprofession; another io or ii years brings them to the peak; andit is then between 15 and 20 years before their incomes fallbelow the average for the profession. The rise is considerably
25 30 35
Years in practice
246 PROFESSIONAL INCOMEsteeper than the decline, and the peak income is received byphysicians in practice about 17 or i8 years.2Strictly speaking,this pattern describes only the relative earnings in a given yearof physicians in practice varying periods. However, the patternfor 1936 is very similar to corresponding patterns for theincomes of California physicians (Table 32 and Chart 20) andthe 1928 gross incomes of physicians in all parts of the coun-try.8 This close similarity among patterns for periods as muchas eight years apart suggests that they can also be interpretedas describing the change through time in the earnings of eachyear's entrants relative to average earnings in the profession asa whole—the earnings 'life cycle', as it were.
While in the early years incomes depend but little on typeof practice, beginning with the 8—i 2 years-in-practice class theaverages diverge considerably. The pattern for general prac-titioners has a peak at about i 5 years, for partial specialists, atabout 25 years. The peculiar pattern f•r complete specialistsmakes it impossible to select any one year as the peak. There isone peak at 20 years, another at 45 years, the last class forwhich we have data. Fluctuations that might arise from chanceare so great. that either the decline from 25 to 35 years or therjse thereafter might be spurious, although the smoothness of2 These statements are more precise than Chart 22 alone would justify, sinceaverages for five-year periods are plotted on the chart. They are based on ourexamination of the annual averages, which suggests that the curve relatingaverage income to number of years in practice crosses the average for the pro-fession as a whole for the first time between and 8 years and for the secondtime at about or years, and reaches a peak at about 17 or i8 years.3 See Leven, Incomes of Physicians, pp. 114. Interestingly enough, therise between 30 and years in practice in the pattern from our data is matchedby a similar rise in the pattern from the California 1933 data. There is no suchrise in the pattern from the 1928 data. An examination of the averages foreach year in practice rather than for the five-year intervals used in Chart 20,reveals that two of the annual averages in the 28—32 years.in-practice dass arelower than any in the years-in-practice class, and two of the latter arelarger than any in the former group. The remaining three in the two, groupsare about equal. A rough test of the significance of the difference between thetwo five-year averages indicates that the observed difference would be exceededin well over one-fifth of random samples. In view of these results and of theabsence of any reasonable explanation for such a kink' we are disposed toattribute the coincidence between the California and our data to chance.
DETERMINANTS OF PROFESSIONAL INCOME 247
the decline and subsequent rise argues against such an inter-pretation.4 In any event, the pattern for complete specialistshas a much sharper upward 'tilt' than the pattern for partialspecialists, and the latter than the pattern for general practi-tioners. Consequently, the percentage differences among thethree groups of practitioners display a fairly uniform tendencyto increase with number of years in practice. These patternsfor the three groups of practitioners cannot, like the patternfor all physicians, be interpreted as describing the 'life cycle'of earnings. Young men who become physicians ordinarily re-main physicians throughout the whole of their working life.Young men who start as general practitioners, however, mayoften later become partial specialists or complete specialists.The patterns for the different groups of practitioners natu-rally reflect this shift in their composition, and hence cannot beinterpreted as the 'life cycle' of the earnings of a stable group.(See also Sec. below.)
Chart 23 presents for each type of practitioner and each sizeof community the relation between average income and yearsin practice. The averages are given in Table The tendencynoted in the countrywide data for the tilt of the patterns toshift upward as we pass from general practitioners to partialspecialists to complete specialists persists for the larger com-munities, but is, if anything, reversed for the two smallestsize of community classes. In general, the amplitude of thepatterns increases at first as size of community decreases, and4 It is difficult to make an accurate test of these statements: first, because theparticular movements to be investigated are selected from many on the basisof their peculiarity; second, because the averages are weighted. However, stand-ard errors of the differences can be computed that neglect these difficulties.The difference between the average at 20 years in practice and the average at30 years is about twice its approximate standard error; the difference betweenthe latter average and the average at 45 years in practice is slightly more thantwice its approximate standard error; and the difference between the averageat 20 years and the average at years in practice is about one and a halftimes its standard error. If we disregard the two complications mentioned,these differences are on the borderline of 'significance'. Since both complica-tions tend to lessen the 'significance' of the differences, it is not unreasonable toattribute the differences to chance fluctuation. It should be noted that thesetests do not take into account the regularity of the averages.
248 PROFESSIONAL INCOME
CHART 23
Arithmetk Mean Income of Phystcians by Type of Practice,
Size of Community1 and Years in Praclice
All sizes ofArithmetic mean income(thousands of dollars)12
I0
8
6
4
2
10 20
Communilies with populations of 500,000 and over
'U
8
6
0 I
I
12
I0
60
50,000—100,000
All phy,ici.as
Arithmetic mean income(thousands of dollars)
General
l0
a
l0
6)
tO 20 30 40 50 I - 10 20
Communities with populations of 100,000—500,000
30 40 50 6
12
10
0 l'02'030405'0 6 -
Communities with populations of
14
12
10
8
6
4
2
0 10 20 30 40 50 60
Years in practice Y..rs in practice
The horizontal line in each diagram represent, the arithmetic mean income of all phyd.din. of the ,ped6ed type In the particular size of a,mmunlty.
DETERMINANTS OF PROFESSIONAL INCOME 249
ChART 23 (coNT.)
Pirisil
of'2
a
6
4
2
i0
a
6
4
2
0
All sizes of communities
mean income(thousands ot doltars)12
10
8
6
4
2
0
Communities wtth populations
I0
8
6
4
2
50 600)
'0
Communities with populations of 100,000 — 500,000
12
l0
8
6
4
2
'4
'2
8
6
4
2
0
Complete
40 61
of 500,000 and over
10 20 30 40 I 61
I 10 20 30 40 50 61
Communitks with populattans of 50,000— 100,000
16
50 60
Y..rs in practice in practice
The horizontal line In each diagram representa the arithmetic wean Income of all phyul.dana of the spedfied type In the particular the of community.
250
CHART 23 (coN'r.)
PROFESSIONAL INCOME
All physiciansCommunities with populations of 25,000— 50,000
General practitioners
Communities with populations of 10,000—25,000
12
I0
8
6
4
2
0 00 tO 20 30 40
Communities with populations of 2500—10,000
12
I0
8
6
4
2
0
6
4
2
10 20 30 40
6
4
2
50 6) 10 20 30 40
Communities with populations under 2,500
6
00 0 4'O 60Years in practice
50 I
4.a
00
The horizontal line in each diagram represents the arithmetic mean Income of all physi-clans of the specified type in the particular size of community.
30 40 50 60
Y,.rs in practici
DETERMINANTS OF PROFESSIONAL INCOME 251
CHART 23 (coNcL.)
tO 20 30 40 50
JoVia,s In practice
The horizontal line in,each diagram represents the arithmetic mean Income of all physi-dana ot the specified type In the particular size of community.
Partial Complete 3pcciaIistspopulatons of 25,000 — 50,000
mean incomeof dollers)
Communities with of l0,000
6
12
to
6
4.
0 io so
CommunIties eith populations of 2,500—10,000
6
- 2'
60 00I I I
CotTImunfties sith populoticns 2,500
2
A
I-U
4
2
0
20 30 40 50 60
to 20 40 50 60
TA
BL
E 3
4
Ari
thm
etic
Mea
n In
com
e an
d N
umbe
r of
Per
sons
Cov
ered
, by
Yea
rs in
Pra
ctic
e an
d Si
ze o
f C
omm
unity
a A
llP
hysi
cian
s,19
36
1,27
02,
541
3,17
94,086
3,914
5,58
85,465
5,359
2,61
35,
470
5,56
25,
858
9,67
713
,000
Ari
thm
etic
Mea
n In
com
e (d
olla
rs)
4,73
34,
547
3,78
14,
503
6,61
53,
985
7,05
44,
860
7,12
23,
881
3,03
98,
121
5,22
04,
599
3,07
45,
014
6,og
g4,
775
3,11
94,
134
3,87
33,060
2,96
83,
232
2,85
32,
211
2,82
1
Nur
n be
r of
Per
sons
Cov
ered
'
Ari
thm
etic
Mea
n In
com
e (d
olla
rs)
3,84
43,
033
2,71
03,
048
1,75
74,
551
4,85
41,
878
2,88
89,
658
2,35
22,
004
i,6ig
2,46
56,
048
5,45
34,404
2,45
92,
231
to to
61
o4 3
21
SC
O6
21
62
I
510
20 z
83i8
7
z C,, 0
200
t'J
NO
T R
EP
OR
T-
SIZ
E O
FIN
GU
nder
CO
MM
uNrF
YA
LL
IN P
RA
CT
ICE
3
500,
000&
OV
eT3,
525
3,74
31,
573
2,35
74,
594
100,
000—
500,
000
5,10
55,
155
1,91
53,
685
5,99
15,
512
50,0
00-1
00,0
005,
079
...2,
166
2,99
86,
277
8,41
625
.000
- 50
,000
4,67
92,
086
1,11
03,745
4,819
10.0
00-
25,0
004,
256
1,674
3,49
24,
715
2,50
0- 1
0,00
03,
648
2,556
3,830
4,805
4,11
1U
nder
2,50
02,
620
2,79
42,
135
3,18
32,
731
5,568
RE
POR
TIN
G Y
EA
RS
IN P
RA
CT
ICE
AS
3—7
8—u
15—
1718
—22
29—
2728
—32
33—
3738
—42
U.S
.3,
9411
4,01
1
43—
4748
—52
55—
57
1,78
85,
182
4,71
85,
326
5,17
94,
8591
3,57
5
3,05
44,
342
7,34
03,
228
2,01
53,
194
1,65
3
500,
000&
over
241
7i8
4210
0,00
0-50
0,00
026
515
2232
50,0
0010
0,00
097
...6
1225
,000
- 50
,000
110
4i6
10,0
00-
25,0
0016
26
502,
500-
10,
000
ig8
97
40U
nder
2,5
008
17
2,41
81,
635
5,47
21,
331
2,82
32,
106
1,65
4
i,6oo 6o
o2,
682
61$
3,76
71.
141
1,27
950
020
070
064
01,
353
4,01
63,
091
2,40
21,
703
747
4226
2229
2826
2612
137
1310
910
iS24
1511
i624
i826
32i8
14
43
U. S
.14
091
48
b G
ener
al P
ract
ition
ers,
193
6
500,
000
&O
ver
2,80
246
010
0,00
0-50
0,00
03,
819
4,160
50,0
00—
100,
000
8,17
125
,000
-50
,000
3,70
12,
250
22i6
1024
3219
8i6
420
97
20i6
1025
1015
4847
35
8i
228
172
125
105
1791
167
146
100
2,73
26o
o1,
250
2,83
91,364
159
05
124
84
51
75
14
149
28
179
105
4522
99
ii6
54
57
79
72
22
31
1
66
45
*
37
85
314
125
84
2931
2850
4
...z
... ..;z
I I *0
20,0
00—
25,
000
2,92
54,
935
2,50
0—10
,000
5,17
41,
933
Und
er 2
,500
2,38
1
U.
S.2,
917
3.26
8
500,
000
&ov
er75
I20
0,00
0-50
0,00
073
750
,000
-100
,000
2125
,000
- 50
,000
412
10,0
00-
25,0
0062
1
2,50
0- 1
0,00
093
3U
nder
U.S
.60
319
c P
artia
lSpe
cial
ists
, 193
6
1.66
32,
574
4,12
85,
231
1.83
25,
957
4,80
53,
454
1,74
42,
725
3,19
54,
454
4,33
42,
211
2,66
22,
903
3.10
74,
875
1,89
05,
473
2,32
92,
530
2,75
42,
584
1.93
82,
087
1,67
5
1,77
03,
114
4,07
13,
500
2,60
52,
538
2,25
7
Num
ber
oJ P
erso
iu C
over
ed'
I 4 1 5 5 25
1,42
254
070
021
22...
1,91
0gg
81,
353
.1
1,96
71,
169
69*
3711
4
1 1 1 9
3,27
64.
595
4,09
95.
534
4,39
14,
165
3,12
7
3717
*71
6*19
125
4,11
66,
405
2,81
7$3,
254
5,20
52,
145
3,02
11,
924
2,05
02,
433
1,70
92,
200
4.67
5
1,58
43,
404
2.90
03,
837
4,37
14,
434
500,
000
&ov
er20
0,00
0-50
0,00
050
,000
-100
,000
25.0
00—
50,
000
10,0
0025
,000
2,50
0- 1
0,00
0U
nder
2,50
0
U.S
.
& o
ver
100.
000-
500.
000
50,0
00-
200,
000
*5.0
00-
50,0
0010
,000
- 25
,000
2,50
0-10
,000
Und
er 2
,500
Ari
thm
etic
Mea
nIn
com
e(d
ofla
rs)
3,4*
44,
832
3,11
84,
530
4,35
13.
375
3,53
82,
122
i,6oo
5,2*
64,
640
6.48
75,
708
4,25
89,
432
2,71
71,
414
...7,
346
5,97
43,
35*
9,84
14,
882
7,68
36,
778
6,,*
*3,
074
4,27
73.
2 15
3,46
16o
28,7
60...
i,o88
2,52
5...
5,57
05,
156
4,18
53,
950
4,49
85,
598
5,50
76,
365
3,41
82,
702
2,85
82,
154
4,22
52,
081
700
3,76
13,
484
5,34
94,
131
2,60
*2,
816
3,93
31,
471
1,00
02,
000
3,95
71
2
048
382
6i3
885
053
1,27
950
0
0 "1 (6'i 0 z z C) 0
1,72
25.
085
4,57
14,
908
4,74
43,
980
4,28
*2,
766
2,40
39*
2
Num
ber
of P
erso
ns C
over
ed'
615
1913
98
107
55
1
910
106
14.
1011
92
...1
24
46
45
28
1...
11
*5
33
54
75
11
113
97
37
55
45
1
28
89
123
5*
2
210
109
513
1719
51
1•.
.30
126
aU
. S.
5131
*8*4
7668
5*43
611
U.S
.29
511
I In
clud
es o
ne r
etur
n fo
r w
hich
size
of c
omm
unity
was
unk
now
n.'N
unib
er o
f pe
rson
s co
vere
d by
the
retu
rns
used
, bef
ore
wei
ghtin
g.
Num
ber
of P
erso
nsy1
8so
64
1
914
712
33
64
52
2s
67
..'3
68
s2
51
22
TA
BL
E 3
4 (c
onci
.)
d C
ompl
ete
Spec
ialis
ts, 1
936
SIZ
E O
FC
OM
MU
NIT
Y
500,
000
& o
ver
4,70
54,
478
100,
000-
500,
000
6,41
65,
515
50,0
00-1
00,0
007,
292
...25
,000
- 50
,000
4,88
5...
so,o
oo-
25,o
oo6,
729
3,00
02,
500-
10,
000
2,71
784
8U
nder
2,5
005,
875
...
Und
er3
2,04
72,
109
1,93
472
6
Ari
thm
etic
Mea
n In
com
e (d
olla
rs)
NO
rRE
FOR
T-
RE
POR
TIN
GY
EA
RS
INPR
AC
TIC
EA
SIN
C Y
EA
RS
AU
.IN
PR
AC
TIC
E5—
78—
is13
—17
18—
2223
—27
28—
3233
—37
38—
4243
—47
3,17
93,
759
3,15
95,
654
3,30
3
4,05
06,
750
7,08
311
,216 ...
4,94
95,
930
13,1
526,
047
5,47
6 ... ...
7,13
79,
154
4,82
717
,657
6,94
75,
053 ...
5,57
49,
100
8,6g
b4,
272
6,51
12,
658
7.50
0
5,71
65,
859
5,41
55,
241
3,54
21,
000 ...
9,38
77,
859
3,84
9 ...5,
494
3,13
4 ...
6,co
o11
,034
14,1
552,
762
2,41
32,
729 ...
... ...10
,749 ... ... ... ...
5,79
5.4,
769
i,8g6
5,54
07,
312
7,49
94,
558
7,15
59,
085
10,7
49
48-5
25r
57
U.S
.
500,
000
&ov
er10
0,00
0—50
0,00
050
,000
-100
,000
25,0
00-
50,0
0010
,000
-25
,000
2,50
0-10
,000
Und
er 2
,500
684
512
144
p85
910
1415
34
36
51...
34
25
39i
...6
6'7
1...
42
2...
...I
so41
2846
2612
2
0 psi
0) 0) 0 z
The
se n
umbe
rs, t
here
fore
, can
not b
e us
ed d
irec
tly to
com
pute
per
-z
cent
age
dist
ribu
tion
of p
hysi
cian
s by
yea
rs in
pra
ctic
e.
0
• DETERMINANTS OF PROFESSIONAL INCOME 255
then declines sharply. Little or no importance can be attachedto the initial increase: the• patterns for the very large com-munities are based on more returns than those for communi-ties of an intermediate size; and the differences in amplitudemay reflect merely the greater play of random variation in thelatter. The lower amplitude of the patterns for small com-munities may reflect the larger number of returns in suchcommunities for all groups except complete specialists. Butthe decline in amplitude is so large that it probably cannot beset aside.
Except for the lower amplitude of the patterns for smallcommunities, the most interesting feature of Chart 23 is thefairly persistent but far from uniform tendency for the pat-terns to assume a greater downward tilt as size of communitydecreases; i.e., for the decline in income to become steeperrelatively to the initial rise. In general, the smaller the com-munity the sooner is the average income of all physicians sur-passed, and the peak income received. For all physicians, thepeak income is at 20 years in the two largest size of communityclasses,5 at i 5 years in the next two, at 10 years in the next two,but at 15 years in the smallest communities. For general prac-titioners, the progression is somewhat less regular: 15, 10, 10,15, 15 (or 25), 10,6 5. For partial specialists and complete spe-cialists, the extreme irregularity of the patterns makes the se-lection of single peaks difficult, but a similar general tendencyis noticeable, particularly for complete specialists.
The difference in the 'tilt' of the patterns is reflected insmaller differences among communities of varying size in theaverage incomes of younger practitioners than in the averageincomes of older practitioners. Indeed, for the first two years-in-practice classes in Table average incomes are highest incommunities with populations of 2,500—10,000 and are notmuch lower in the smallest communities. Nonetheless, theranking of the size of community classes by average income is
5For communities of this assumes that the peak at years isto be disregarded.6 Disregarding peak at 40 years.
256 PROFESSIONAL INCOMEmuch the same for most of the years-in-practice classes sepa-rately as for all combined. Differences among the size of com-munity classes in the distribution of physicians by years inpractice seem not to be responsible for the size of communitydifferences in average income analyzed in the preceding chap-ter.T
The relatively flat pattern in small communities is prob-ably attributable to the absence of effective direct competition.In large communities there are many physicians and the com-petitive forces discussed above can operate. The young physi-cian only recently out of medical school appeals to some be-cause he is presumably 'up-to-date' or merely because he is'new'; the middle-aged man, to others who regard his 'practicalexperience' as more than counterbalancing the possibilitythat he may not be familiar with the latest methods; the olderman, to still others, whose choice is dominated by habituation,confidence born of long and continuous service, or respect fora reputation gained by many years of practice. The relativeincome status of physicians in practice varying periods dependson the relative strength of these motives; the general prefer-ence for middle-aged men leads to their receiving higher in-comes. In villages, on the other hand, there are frequently onlyone or two physicians. Competition is almost completely in-effective. The income a physician receives depends less on hisage than on other factors—the prosperity of the region inwhich he practises, its proximity to other places, and the like.Of course, there is some competition and hence some tendencyfor income to rise at first with number of years in practice. Also,the older men become partly retired and hence income even-tually declines. But both the rise and decline are mild rela-tively to those in larger communities.
One possible explanation of the tendency for the tilt of thepatterns to shift downward as size of community decreases isthe concentration of young physicians in the larger communi-ties (Table 35). In cities with over a million inhabitants more7 Size of community averages standardized for differences in distribution byyears in practice are almost the same as the original averages.
TA
BL
E 3
5
Age
Dis
trib
utio
n, b
y Si
ze o
f C
omm
unity
All
Phys
icia
ns, G
ener
al P
ract
ition
ers
and
Part
ial S
peci
alis
ts, a
ndC
ompl
ete
Spec
ialis
ts,
i6 S
tate
s an
d th
e D
istr
ict o
f C
olum
bia
SIZ
E O
F C
OM
M 1
3N
IT
YC
LA
SSZ
1,00
0.00
050
0,00
0-25
0,00
0-10
0,00
0-50
,000
-25
,000
-10
,000
-5,
000-
2,50
0-1,
000-
Und
erA
LL
AG
E&
Ove
r1.
000,
000
500,
000
250,
000
100,
000
50,0
0025
,000
10,0
005,
000
2.50
03,
000
CL
ASS
ES
All
Phys
icia
ns0
Und
er28
.527
.1sa
g21
.620
.215
.915
.914
.214
.212
.010
.3*1
.1
35—
4427
.925
.925
.027
.724
.923
.6*3
.020
.518
716
.914
.623
.9
45—
5421
.121
.323
.323
.126
.8*6
.8*5
425
.126
.725
.724
.723
.5&
ove
r22
.525
.7*8
.827
.628
.153
.795
.740
.240
.445
450
.431
.5o
MI
ages
100.
010
0.0
100.
010
0.0
100.
010
0.0
100.
010
0.0
100.
010
0.0
100.
010
0.0
Gen
eral
Prac
titio
ners
and
Par
tial S
peci
alis
tsU
nder
35
534
31.8
264
23.3
16.9
17.3
14.8
14.6
12.0
10.3
23.2
'4
35—
4426
.223
.222
.2*4
.7*1
.521
.8ig
.g18
.416
.714
.421
.90
45—
5418
.7i8
.820
.6*0
.2*5
.325
.424
.224
.326
.225
.624
.522
.0Z
55&
ove
r21
.7a6
.**9
.529
.9S
6.*
41.0
40.8
454
no.8
52.9
All
ages
ioo.
o10
0.0
100.
010
0.0
100.
010
0.0
100.
010
0.0
100.
010
0.0
100.
010
0.0
e
Com
plet
eSp
ecia
lists
ZU
nder
10.9
114
ii.6
8.8
11.0
12.3
849.
17.
89.
28.
gio
.635
—44
34.0
34.6
55.1
54.7
31.5
29.3
25.9
23.8
25.8
22.6
045
5429
.7*g
.651
.552
.151
.531
.731
.731
.333
.335
.1&
ove
r25
.424
.4*4
4no
.633
456
.633
.425
5
All
ages
100.
010
0.0
100.
010
0.0
100.
010
0.0
100.
010
0.0
100.
010
0.0
100.
010
0.0
R.
G. L
elan
d, D
istr
ibut
ion
of P
hysi
cian
s in
the
Uni
ted
Stat
es (
Am
eric
anof
fin
er c
lass
es f
or w
hich
the
orig
inal
dat
a ar
e pr
esen
ted.
Med
ical
Ass
ocia
tion
pp. 5
6-7.
The
age
cla
sses
are
wm
biua
tionz
258 PROFESSIONAL INCOMEthan a quarter of the physicians were under 35 years of agein 1931, while fewer than a quarter were 55 or over; in placesof less than a thousand inhabitants, on the other hand, only atenth of the physicians were under years of age while halfwere 55 or over. The exact proportions are very different forthe two groups segregated—complete specialists, and generalpractitioners and partial specialists—but the general picture isthe same.8
The average income of young physicians will depend notonly on the preferences of the community for physicians inpractice varying periods, but also on the relative number ofyoung physicians. For any given community, the higher theproportion of young physicians, the lower their average in-come will tend to relatively to the average income of oldermen. We have, of course, no evidence on whether or howpreferences for physicians vary with size of community. Butwe do know that the smaller the community the better therelative income status of the young physician. And it seemsreasonable to attribute this to the relative scarcity of youngphysicians in the small communities. Of course, this explana-tion must be related to the preceding discussion.of the degreeof competition in communities of varying size. It can be con-sidered valid only for communities sufficiently large to permita moderate degree of competition, say for communities overio,ooo in population.
Why are the age distributions of physicians so different incommunities of different size? Two explanations may be sug-gested. Perhaps the less important is the increasing need forextensive equipment and adequate facilities that has tendedto make large cities seem more desirable to the physician thanrural communities where facilities are likely to be limited.Probably of far greater importance is the secular increase inthe proportion of the nation's population living in big cities,
8 Distributions by years in practice and size of community computed from oursample suggest the same tendency. But the estimated distributions vary somuch and so irregularly among size of community dasses that the tendencycould not be established from our data alone.
DETERMINANTS OF PROFESSIONAL INCOME 259
largely a result of migration from rural to urban communities.The corresponding shift in the distribution of physiciansamong communities of different size ° probably took place lessthrough actual movement of physicians already in practicethan through concentration in the larger cities of physiciansjust entering practice. The shift in population might be ex-pected to lead to low medical incomes in rural communitiesrelatively to incomes in large communities, and thus providea strong incentive for new physicians to begin practice inlarge communities—indeed, the proverbially low income ofphysicians in rural areas is amply supported by our data. Thesmall number of new entrants who started practice in ruralareas presumably led to a smaller decline, or greater rise, inincomes in those areas relative to incomes in urban communi-ties than would otherwise have taken place. Further, it led toa relatively favorable income status of young physicians, andhence to smaller size of community differences in the averageincomes of young physicians than in the average incomes ofolder physicians.
If our analysis is correct, and the distribution of the popula-tion among communities of different size remains fairly sta-tionary for a moderately long period, or at least changes lessrapidly than in the past, we may expect a lessening of the dif-ference between average incomes in small and large communi-ties, and between the age distributions of physicians incommunities of different size, and as a consequence, a greatersimilarity in the relation between income and years in prac-tice. It follows that while the pattern for all physicians relatingincome to years in practice may be interpreted as describingthe 'life cycle' of earnings, the patterns for the separate size ofcommunity classes cannot be so interpreted.
Differences in age distribution are not, of course, the onlycause of size of community differences in the relation betweenincome and years in practice. Differences in clientele, the vari-ability of physicians' incomes, the opportunity to obtaina see Leland, Distribution of Physicians. pp. 50. 51, evidence that such. a shifthas taken place.
260 PROFESSIONAL INCOMEsalaried posts, and numerous other factors doubtless play arole. Consequently, even if differences in age distribution wereeliminated, the relation between income and years in practicemight still be expected to differ among communities of vary-ing size.
3 TYPE AND ORGANIZATION OF PRACTICE
The increasing complexity of professional activity has led toconsiderable specialization. The accompanying division offunction has taken different forms in different professions. Inmedicine and dentistry it has taken the form of a differentia-tion in the kind of service rendered by individuals practisingindependently and coordinated by the impersonal mechanism
/of..the market place. Sharing of office space, nonprofessionalhelp, and equipment is frequent, but formal organization intopartnerships or firms, exceedingly rare. Few physicians ordentists employ others; only about 20 per cent of all physiciansand dentists are salaried employees and many of these are em-ployed by business and government. In accountancy, engineer-ing, and to some extent law, division of function has taken theform not only of specialization by separate professional units,•but also of larger professional units, the individual membersor employees frequently concentrating on specific fields ofpractice. The larger professional units are sometimes con-ducted by an individual practitioner who hires other profes-sional men as employees; more frequently, they are organizedas partnerships or firms of several independent professionalmen who may or may not employ others. Our data suggest thatabout a third of the accountants, almost half of the engineers,and abciut a quarter of the lawyers who practise independentlyare members of firms.
The form division of function has taken is primarily at-tributable to the type of consumer served and the nature of theservices rendered. The curative professions serve almost ex-clusively individuals in their capacity as ultimate consumers.The 'unit' of professional service is therefore small and can
DETERMINANTS OF PROFESSIONAL INCOME 261
seldom be delegated in large part to subordinates. The indi-vidual customer usually requires only a single type of serviceat any time; he seldom purchases a 'bundle' of different kindsof service. The kinds of service needed can rarely be predictedor contracted for in advance. The necessary professional equip-ment can ordinarily be almost as efficiently and intensivelyutilized by one man as by many, or by a sharing arrangement'that does not affect other aspects of practice. Since prestige andjudgment of quality attach to the individual practitioner, littleis to be gained by organization into 'firms', or by the hiringof many professional employees.
If hospitalization had developed as a private enterprise un-der the direct control of physicians, instead of as a semipublicenterprise under the control of nonprofit or governmental or-ganizations, these statements would not be valid. The elabo-rate physical equipment and the large number of subordinatesthat would then have been needed might have led to a verydifferent organization of medical practice than we now have.Firms and even incorporated enterprises might have becomecommon. Whatever private hospitals do exist are frequentlyconducted by 'firms' of physicians.
The increasing complexity of medical practice and the in-creasing emphasis on preventive care and periodic physicalexaminations are giving rise to a greater need for extensiveequipment as well as a greater possibility of cooperationamong physicians in rendering medical services. In addition,the widespread agitation for some form of health insurance orcooperative purchase or provision of medical care is tendingto enlarge the 'unit' of service purchased and to facilitate ad-vance contractual arrangements for an assortment of special-ized services. Both factors seem to be leading to an extension ofthe 'group' practice of medicine, and, to some extent, of den-tistry as well.
Law, to a significant degree, and accountancy and consult-ing engineering almost exclusively, render services to businessenterprises. The 'unit' of professional services tends to be
262 PROFESSIONAL INCOMElarger and can more easily be delegated to subordinates thanin the curative professions. In addition, individual customersare more likely to purchase a variety of services and can moreeasily predict and contract for them in advance. Organizationinto firms seems to be least frequent among lawyers, theonly one of the three 'business' professions that renders a sig-nificant part of its services to individuals.
The difference in clientele that accounts for the difference inthe form that division of function has taken and for the conse-quent difference in the size of professional units has also af-fected another aspect of the organization of professionalactivity, namely, the relative importance of independent andsalaried practice. As we have seen, few independent physiciansor dentists hire others as salaried employees. In addition, fewusers of medical or dental service need or can afford a full-timephysician or dentist. Only when business enterprises or gov-ernmental bodies provide medical care for their employees orwhen government engages in public health activities is therean opportunity for salaried employment by nonprofessionalenterprises; and even in these cases, the part-time services of aphysician or dentist engaged principally in independent prac-tice often suffices. As a result, about 8o per cent of all physi-cians and dentists are in independent practice. In the businessprofessions, on the other hand, not only is salaried employ-ment by professional units proper more common, but also theprincipal consumers are business enterprises and govern-mental bodies that can afford to hire full-time professionalemployees, and that use the independent professional unitsonly for specialized services their own employees cannot ren-der. Over nine-tenths of all engineers and accountants andauditors (certified and noncertified) and about a third of alllawyers are salaried employees.
a Type of practiceSpecialization is far more widespread in medicine than indentistry. Of the physicians listed in the i 931 Directory of the
DETE1tMINANTS OF PROFESSIONAL INCOME 263
American Medical Association, i6 per cent considered them-selves complete specialists, i6 per cent, partial specialists, and68 per cent, general practitioners.1° According to our 1937medical sample, the only one of our samples in which therespondent was asked type of practice," 22 per cent of thephysicians considered themselves complete specialists, 38 percent, partial specialists, and 40 per cent, general practition-ers.12 The study of the Committee on the Costs of MedicalCare, on the other hand, indicates that in 1929 only 3 per centof the dentists considered themselves complete specialists, 8per cent, partial specialists, and 89 per cent, general practi-tioners.18 Unfortunately, no data are available on the trend ofspecialization in dentistry. While in this, as in other respects,dentistry is likely to develop In the same manner as medicine,but with a considerable lag, it is doubtful that specializationwill ever be as widespread in dentistry as in medicine, since,
loLeland, Distribution of Physicians, p.11 The actual wording of the question was:
"Indicate the type of practice in which you are engaged:General practiceSpecialized practice (such as surgery, neurology, obstetrics, etc.)Special interest with general practice (such as pediatrics, surgery, gyne-
cology, etc.)12 The percentage of complete specialists in our sample checks fairly doselywith other estimates, but the percentage of partial specialists is considerablyhigher (see Table and Appendix A, Sec. 2C i). This difference does not neces-sarily reflect bias in the sample. It may more reasonably be interpreted as aresult of differences in the way the question on specialization was asked andthe answers were edited, since the distinction between general practitionersand partial specialists is exceedingly vague. There are no commonly acceptedobjective criteria to segregate general practitioners from physicians giving'special attention' to a subject but not limiting their practice to it. In mostcases, therefore, physicians are asked to dassify themselves. In a survey forMichigan, however, an attempt was made to go behind this classification byasking the public relations committee of each county medical society to classifythe society's members. The results show that a considerable percentage of thephysicians who classified themselves as partial specialists were considered bytheir colleagues to be general practitioners. See Report of the Committee onSuruey of Medical Services and Health Agencies (Michigan State Medical So-ciety, 1939),18 See Table 36, which gives also the estimates of the California Medical-Eco-nomic Survey. The two gets of estimates are very close.
264 PROFESSIONAL INCOMEin a very real sense, dentistry is itself a specialized branch ofmedicine. In medicine the available data suggest, though theevidence is by no means conclusive, that specialization is stillincreasing.'4
In both professions, complete specialists receive higher aver-age incomes than partial specialists, and partial specialists thangeneral practitioners. Though the differentials cannot beestimated with any exactitude from the studies summarized inTable they are sizable. In dentistry, the average income ofcomplete specialists seems to be about 30 per cent larger thanthat of partial specialists, and the average income of partialspecialists about per cent larger than that of general prac-titioners. In medicine, for the period 1929—36, according toour 1937 sample, the average income of complete specialistswas approximately $5,900; of partial specialists, $3,800; ofgeneral practitioners, $2,900; i.e., the average income of com-plete specialists was about 50 per cent larger than that ofpartial specialists, and the average income of partial special-ists, about 30 per cent larger than that of general practitioners.These estimates of the differentials in medicine are in generalbetween those suggested by the other two studies of medical in-comes cited in the table: the estimates from the study of theCommittee on the Costs of Medical Care are higher, from theCalifornia survey, lower. Although based on fewer returns,the estimates from our 1937 medical sample are probably more14 See Leven, Incomes of Physicians, pp. 50_65; H. G. Weiskotten,in Medical Practice', Journal of the Association of American Medical Colleges,March 1932, pp. 65—85.
Dean Weiskotten's study gives data on the type of practice in 1926 of 1,834physicians who graduated in 1915. and 1,947 who graduated in 1920 and on thetype of practice in 1931 of 3,230 physicians who graduated in 1925. The datasuggest that the proportion of graduates who became or who planned to becomespccialists was fairly stable for the classes of 1920 and 1925, and greater forboth than for the class of 1915 or all physicians. The increasing specializationamong physicians as a whole seems to reflect the replacement of older phy-sicians by younger, associated with a relatively stable degree of specializationamong the younger. The evidence on this point, however, is fairly meagre.
Our 1937 medical sample provides no information on changes over time in thedegree of specialization, since the questionnaires requested merely the type ofpractice when the questionnaire was received (early 1937).
DETERMINANTS OF PROFESSIONAL INCOME 265
TABLE 36
Arithmetic Mean Income and Percentage Distribution, by Typeof Practice
Physicians and Dentists: Selected Studies
DENTISTSPHYSICIANS Conim.on
Comm. on Cal. Costs of Cal.Costs ol Medical. Medical Medical-
193'7 Medical Economic Care'(2o Economicsample1 Care2 Survey8 states) Survey8'1929—36 1929 1929-34 1929 1929—34
Arithmetic mean income (dollars)General practitioners 2,922 s,goo 3,836 4,791 3,564Partial specialists 5,777 6,ioo 5,060 6,129 4,847Complete specialists 5,904 10,000 6,039 8,625 5,894
All practitioners 5,700 4.916 5,011 3,762
excess ofPartial specialists over general
practitioners 29 56 32 28 36Complete specialists over general
practitioners 102 156 57 8oComplete specialists over partial
specialists 64 19 41 22
% of practitioners who were 8General practitioners 40 42 89 90Partial specialists 8 5Complete specialists 22 23 84
All practitioners 100 100 100 100 100
No. of returns g,7377 4,705
See Tables 37. 41, and Percentage distribution is for early 1937.Maurice Leven. Incomes of Physicians (University of Chicago Press, 1952), pp. ig, 15,
109.8 California Medical-Economic Survey, 73, 79, 94, 100. Percentage distributions are for1934.'Leven, Practice of Dentistry, pp. 14, 88, 89.
Number of persons reporting 1956 income, before weighting.C Estimated number of returns received, before weighting or adjusting; computed fromLeven, Incomes of Physicians, pp. 13, 15, 109.
Number of persons reporting 2953 incomes; fewer reported for other years. Percentage dis-tribution based on 3,206 returns for physicians and 1,802 for dentists.
Distributions are based on the number reporting in the last-year of the period Indicatedabove.
TA
BL
E 3
7
Ari
thm
etic
Mea
n In
com
e, b
y T
ype
of P
ract
ice
Phys
icia
ns, 1
929—
1936
Ave
rage
1929
1930
1931
1932
193)
5935
1936
1929
—36
Arit
hmet
icm
ean
inco
me
(dol
lars
)G
ener
al p
ract
ition
ers
4,09
63,
571
3,06
724
042,
266
2478
2,58
42,
914
2,92
2Pa
rtia
l spe
cial
ists
5,33
74,
467
4,01
53,
018
2,78
83,
215
3412
3,96
23,
777
Com
plet
e sp
ecia
lists
8,29
07,
787
6461
4,69
44,
187
4,83
65,
174
5,80
55,
904
exce
ssof
Part
ial s
peci
alis
ts o
ver
gene
ral p
ract
ition
ers
30.5
25.1
50.9
25.5
23.0
29.7
52.0
56.0
29.5
'C
ompl
ete
spec
ialis
ts o
ver
gene
ral p
ract
ition
ers
1024
118.
111
0.7
95.3
84.8
95.2
100.
299
.210
2.1'
Com
plet
e sp
ecia
lists
ove
r pa
rtia
l spe
cial
ists
55.5
74.5
60.9
55.5
50.2
504
51.6
46.5
56.3
'
No.
of
retu
rns
2G
ener
alpr
actit
ione
rS38
235
937
940
143
453
155
40
Part
ial s
peci
alis
ts34
232
533
736
545
046
851
2C
ompl
ete
spec
ialis
tsi8
818
319
020
722
125
627
229
3
tCom
pute
d fr
om th
e av
erag
es f
or 1
929—
36.
'of
retu
rns
rece
ived
, bef
ore
wei
ghtin
g.—
DETERMINANTS OF PROFESSIONAL INCOME 267
accurate than those from the other studies.'5 (The annualfigures from the 1937 medical sample are given in Table
As might be expected from these differences in income level,
TABLE 88
Median and Quartile Incomes, by Type of Practice
Physicians, 1934—19561931 '935 1936
(dollars)
Third QuartileAll physicians 4,290 4,473 5,056General practitioners 3,436 3472Partial specialists 4,338 4441. 5,066Complete specialists 6,i i3 6,902 7,775
MedianAll physicians 2,690 2,824 3,100General practitioners 2,221 2,220 2,484Partial specialists 2,749 2,900 3,230Complete specialists 3,798 4,165 4,575
First QuartileAll physicians 1,554 1,824General practitioners 1,234 1,251 1,394Partial specialists 1,798 i,8o6 2,014Complete specialists 2,238 2,329 2,599
absolute variability of income is much greater for completespecialists than for partial specialists, and for partial special-ists than for general practitioners (Tables 38—40). The greaterpart of these differences in absolute variability is accounted
15The estimates from the Committee on Costs of Medical Care study wereobtained indirectly from data on gross income; the relation between gross andnet income was derived from a study of 1928 incomes made by the AmericanMedical Association; see Leven, Incomes of Physicians, p. log. This indirectprocedure might easily have introduced substantial errors. Not only are theCalifornia data for a single state, but also the percentage of complete specialistsSeems too high. Only 20 per cent of California physicians were listed as com-plete specialists in the 1931 Directory of the American Medical Association(Leland, Distribution of Physiciari.c, p. '7), and only 25 per cent of the returnsfrom California in our 1937 medical sample were from complete specialists.The relatively low percentage differences between the average incomes fromthe California study for complete specialists and the other two groups mayreflect the inclusion of some individuals as complete specialists who should havebeen considered partial specialists or general practitioners.
268 PROFESSIONAL INCOMEfor by the differences in income level. Relative variability isin general somewhat less for general practitioners than for theother two groups, but the difference is neither large nor con-sistent.
Complete specialists naturally tend to be concentrated in
TABLE 39
Measures of Variability of Income, by Type of Practice
Physicians, 1934—19361934 1935 1936
lnterquartileDifference (dollars)All physicians 2,736 2,885 3,232General practitioners 2,202 2,221 2,574Partial spedalists 2,540 2,635 3,052Complete specialists 3,875 4,573 5,176
Standard Deviation (dollars)All physicians 2,965 3,057 3,631General practitioners 1,857 1,996 2,242Partial specialists 2,451 2,644 5,572Complete specialists 4,455 4,308 4,831
Relative lnterquartile DifferenceAll physicians 1.017 1.022 1.043General practitioners .991 1.000 1.036Partial spedalists .924 .909 .945Complete specialists 1.020 1.098 1.183
Ratio of QuartilesAll physicians 2.761 2.817 2.772General practitioners 2.784 2.775 2.846Partial specialists 2.413 2459 2.515Complete specialists 2.731 2.964 2992
Coefficient of VariationAll physicians .886 .862 .902General practitioners .733 .755 .752Partial specialists .789 .758 .882Complete specialists .8i6 .817
the larger communities where the more extensive 'market' forprofessional services affords greater opportunity for divisionof function. In addition, specialists are more likely than gen-eral practitioners to be at the peak earnings stage of their work-ing life. Many men begin their careers as general practitionersand become specialists only after receiving additional training
DETERMINANTS OF PROFESSIONAL INCOME 269
and experience. On the other hand, most of the older men re-ceived their professional education before specialization hadprogressed very far and when formal training was less exten-sive than it later became. These differences between specialistsand general practitioners help to explain the large differencesbetween their average incomes. In effect, countrywide aver-ages by type of practice compare one group in the best loca-tions—at least in money terms—and in the prime of life with
TABLE 40
Median and Quartile Incomes, and Measures of Variability,by Type of Practice
Dentists in 20 States, 1929:Committee on Costs of Medical Care Study
ALL GENERAL PARTIAL COMPLETE
DENTISTS PRACTITIONERS SPECIALISTS SPECIALISTS
Third quartile (dollars) 6,203 5,955 7,222 11,765Median (dollars) 4,094 4,756 6,g8oFirst quartile (dollars) 2,669 2,618 5,031 4422
Interquartile difference(dollars) 3,534 3,337 4,191 7,343
Standard deviation(dollars) 4,080 3476 5,930 6,781
Relative interquartiledifference .863 .837 .88i 1.052
Coefficient of variation .8o8 .728 .952 .764
No. of returns 4,705 .4.189 377 139
Computed from frequency distributions in Leven, Practice of Dentistry, p. 88.
another containing many who are in poorer locations and whoare just getting started or are on the verge of retiring.
Limitations of data make it necessary to restrict furtheranalysis of these factors to physicians. Both Table i, based onour sample, and Table 42, based on a comprehensive count ofthe 1931 Directory of the American Medical Association, in-dicate that the percentage of all physicians who are specialistsvaries markedly among communities of different size. Accord-ing to our sample, slightly over 38 per cent of all physicians incommunities with populations of 50,000—500,000 considered
270 PROFESSIONAL INCOME
TABLE 41Distribution by Type of Practice,by Size of Community and by RegionPhysicians, 1936
OF PHYSICIANS WHO WEREGeneral Partial Complete
of community practitioners specialists specialists500,00o & over 30.8 41.6 27.6100,000—500,000 25.2 36.3 38.550,000—100,000 24.8 37.1 38.125,000— 50,000 38.0 37.8 24.210,000— 25,000 41.7 38.0 20.32.500- 10,000 44.8 50.1 5.1Under 2,500 71.9 27.0 1.0
RegionNew England 40.3 36.8 22.9Middle Atlantic 33.0 44.0 23.0E. N. Central 48.7 32.8 18.5W. N. Central 47.8 28.3 24.0S. Atlantic 89.0 37.8 23.2E. S. Central 48.7 15.3W. S. Central 32.6 42.2 25.2Mountain 44.5 37.4 i8.iPacific 33.1 43.2 23.6
U.S. 40.2 58.0 21.9
TABLE 42Distribution by Type of Practice, by Size of CommunityPhysicians Listed in 1931 Directoryof American Medical Association
% OF PHYSICIANS WHO WEREGeneral practitioners Complete
Size of community & partial specialists specialists1,000,ooo&over 79.1 20.9 223
500,000—1,000,000 74.7 3 25.3
250,000— 500,000 74.9 75 325.1
2100,000— 250,000 75.8 5 24.2
50,000— 100,000 76.3 76.3 23.7 23.725,000— 50,000 79.8 79.8 20.2 20.210,000— 25,000 84.0 84.0 i6.o i6.o
5,000— 10,000 88.8 91.2 11.28.8
2,500— 5.000 93.6 6.4 3
1,000— 2,500 97.2 2.82Under i,ooo g8.o 5 2.0 5
Leland, Distribution of Physicians. p.
DETERMINANTS OF PROFESSIONAL INCOME 271
themselves complete specialists while only 25 per cent con-sidered themselves general practitioners. In communities un-der 2,500 on the other hand, only i per cent consideredthemselves complete specialists, while over 70 per cent consid-ered themselves general practitioners. Perhaps the most sur-prising feature of the tables is the lower percentage of completespecialists in the largest communities than in communities ofan intermediate size.1° For the other classes, the percentage ofcomplete specialists declines consistently with size of com-munity.
Size of community differences in the average income of each
18 The lower percentage of complete specialists in the largest communities isnot accounted for by differences among the size of community classes in thedistribution of physicians by number of years in practice; for each years-in-practice class separately, the percentage of complete specialists tends to belower in the largest communities than in those of intermediate size.
To test this point we computed from the 1937 sample the percentage of allphysicians who are complete specialists for each size of community and 9classes by number of years in practice (in practice 2 years or less, to 7 years,8 to 12 years, . . . , to 42 years). In 7 of the 9 years-in-practice classesthe percentage of complete specialists was less in communities over 5oo,oooin population than in either of the next two size of community classes. Theexceptions were the 8—12 and 23—27 years-in-practice classes. Standardized per-centages, using as weights the distribution of all physicians in the United Statesby number of years in practice, were also computed. They differ only slightlyfrom the percentages in Table 41.
More extensive data for i6 states and the District of Columbia on the agedistribution of physicians in by type of practice and size of communityyield a similar conclusion (Leland, Distribution of Physicians, pp. 56—7). Dataare given for two groups: 'general practice or special attention' and 'specialists'.The latter group may be taken to correspond to our 'complete specialists'.Data are given for i 2 five-year age dasses between 20 and 8o (20—24, 25—29,etc.) and for a final class of 8o and up, and for the size of community dasses inTable 42. In 6 of the 12 age dasses for which there were complete specialists inall or almost all size of community classes, the percentage of complete special-ists in communities over 500,000 was less than in communities between ioo,00oand 500,000 or between and ioo,000; in only three classes was the per.centage of complete specialists in communities over 500,000 greater than ineither of the other two size of community dasses. Standardized percentagesusing the age distribution of all physicians in the country as weights differedlittle in magnitude or order from the original percentages. It is significant,however, that the three age classes in which the largest communities had thehighest percentage of specialists included the older physicians, the age classesbeing 55—59, 65—69, and
TABLE 43
Arithmetic
Mea
n In
com
e, R
elat
ives
of Arithmetic Mean Income, and Number of Persons Covered,
by
Size
of
Com
mun
ity a
nd b
y R
egio
n
All
Phys
icia
ns, G
ener
al P
ract
ition
ers,
Par
tial S
peci
alis
ts, a
nd C
ompl
ete
Spec
ialis
ts
AR
ITH
ME
TIC
ME
AN
IN
CO
ME
,1 1
929-
36A
llG
ener
alPa
rtia
lC
ompl
ete
phys
i.pr
acti.
spec
ial,
spec
ial-
clan
stio
ners
ists
isIs
Size
of c
omm
Uni
ty(d
olla
rs)
500,
000
&over
3,87
53,
230
3,45
95,
268
100,
000—
500,
000
5,02
65,
685
4,31
06,
7o8
50,0
00—
100,
000
4,82
63,
153
3,86
06,
778
25,0
00—
50,0
004,
486
5,35
65,
029
5,27
510
,000
—25
,000
4,02
45,
054
3,85
95,
804
2,50
0- 1
0,00
05,
530
3,3'
773,
711
3,03
2U
nder
2,5
002,
448
2,17
03,
131
3,67
2
New
Eng
land
4,74
53,426
4,40
67,
142
Mid
dle
Atla
ntic
4,06
22,
782
6,04
1E
.N. Central
3,83
25,
147
3,89
05,
602
W.N
.C
entr
al3,
310
2,58
02,
916
5,37
6S.
Atla
ntic
4,23
53,
042
4,64
85,
762
E. S
. Cen
tral
2,476
3,406
4,955
W. S.
Cen
tral
3,38
7z,
8oo
2,98
45,996
Mountain
4,28
45,
029
4,28
77,144
Pac
ific
4,23
13,
919
3,99
65,
166
3,91
62,922
3,77
75,
904
1C
ompu
ted
by a
ddin
g th
e av
erag
es f
or th
e ye
ars
cove
red
by th
esa
mpl
e an
d di
vidi
ng b
y th
e nu
mbe
r of
yea
rs—
8.
RE
LAT
IVE
S O
FA
RIT
HM
ET
IC M
EA
N I
NC
OM
E(u
. S. =
ioo)
All
General
Partial
Complete
physi-
prac
ti-sp
ecia
l-sp
ecia
l-cl
ans
tione
rsis
tsis
Is
99.0
110.
591
.189
.212
8.3
126.
011
4.1
123.
210
7.9
102.
211
4.8
114.6
114.
215
3.1
102.
810
4.5
102.
298
.390
.111
5.6
98.5
51.4
62.5
74.5
82.9
62.2
121.
211
7.2
116.
712
1.0
105.
795
.297
.910
2.3
97.9
107.
710
3.0
94.9
84.5
88.5
77.2
91.1
io8.
i10
4.1
123.
197
.677
.684
.790
.283
.986
.56i
$10
1.6
109.
410
3.7
113.
512
1.0
io8.
o13
4.1
105.
887
.5
100.0
100.
010
0.0
100.
0
9868
9498
38 3831
6i39
8717
952
0 'TI
62
398°
6g91
4798
62
37
88
5044
064
35
Z33
2710
5546
2596
7429
58z
Reg
ion
NU
MB
ER
OF
PER
SON
S C
OV
ER
ED
,2 1
936
All
Gen
eral
Part
ial C
ompl
ete
phys
i-pr
acti-
spec
ial-
spec
ial-
clan
stio
ners
ists
Ists
75 75 21 41 62 93 258
U. S.'
241
265 97 110
162
197
335
131
207
197
155 70 199
i6i
1,40
8512
Num
ber
ofpersons
cove
red
bythe
retu
rns
used
, bef
ore
wei
ghtin
g.$
Incl
udes
i re
turn
for
which
size
of
com
mun
ity w
as u
nkno
wn.
295
c•:i 0
DETERMINANTS OF PROFESSIONAL INCOME 273
of the three groups of physicians are similar in character,though in general somewhat smaller in magnitude, than thecorresponding differences in the average income of all physi-cians (Table 43). The peak income of general practitioners isin the same size of community class as the peak income of allpractitioners—communities with populations between ioo,oooand 500,000. The peak income of partial specialists is in the25,000—50,000 class and of complete specialists, in the 50,000—ioo,ooo class. However, the average income of complete spe-cialists is only a trifle smaller in the 100,000-500000 class thanin the 50,000-100,000 class, and is considerably larger in eitherthan in any other class. In general, average incomes arelowest in the smallest communities. The one exception—com-plete specialists-_is hardly significant in view of the fewcomplete specialists in our sample for the two smallest size ofcommunity classes. The variation in average income from onesize of community class to the next is less regular for each typeof practitioner than for all physicians. Presumably, this, too,is attributable to the fewness of the returns on which theseparate averages are based. As a result, little confidence canbe attached to the exact quantitative differences in the table.
Standardized averages provide a simple and efficient toolfor eliminating the effect of differences in the location of theseveral types of physicians. Table 44 gives standardized aver-ages for general practitioners, partial specialists, and completespecialists for 1929—36. They are weighted combinations ofthe size of community averages, weights being the totalnumber of physicians f all types in the corresponding size ofcommunity class.'7 The use of the same weights for all typesof practice eliminates the influence of size of community.18
17 The total number of physicians in each size of community class was takenfrom Leland, Distribution of Physicians, and is for the latest year forwhich such figures are available.18 The standardized averages do not eliminate the influence of regional varia-tion in the distribution of physicians by type of practice except as such variationreflects differences in the size of community composition of the regions. Experi-mental computations for 1936 indicated that eliminating the influence ofregional variation would affect the averages but slightly and the differences
274 PROFESSIONAL INCOMEThe correction for the influence of size of community
sharply reduces the income differentials. The difference be-tween the original averages for complete specialists and gen-eral practitioners is 102 per cent; for complete specialists andpartial specialists, 56 per cent; for partial specialists and gen-eral practitioners, 29 per cent (Table 37, averages, for 1929—
The corresponding differences between the standardizedaverages are 64, 36, and 20 per cent. Roughly a third of thedifference between the original averages is attributable to theconcentration of specialists in the larger communities. The re-maining two-thirds measures the difference between the aver-age incomes of physicians differing in type of practice butpractising in communities of the same size.
The countrywide standardized averages might of course con-ceal important differences among size of community classes orregions in the relative incomes of the three types of practi-tioners. To test whether there are any consistent differenceswe computed, for each size of community in each region, theratios between the average i 936 incomes of complete special-ists and partial specialists, complete specialists and generalpractitioners, and partial specialists and general practitioners.We found no evidence that these ratios differed significantlyamong communities of varying size or among regions.'9 How-
among the averages even less. These computations were made by weightingthe average income of each type of practitioner in each size of communityand regional 'cell' by the total number of physicians in that as shownby Leland. The averages for the different types of practice were made com-parable by excluding those cells in which there were no returns for one ormore types of practice. The resulting standardized averages are $3,114 forgeneral practitioners, $4,035 for partial specialists, and $5,587 for completespecialists. All are somewhat higher than the averages in Table 44 because theexcluded cells induded mainly the smaller communities. According to theseaverages the income of complete specialists exceeds that of general practitionersby 79.4 per cent, and of partial specialists by 38.5 per cent, while the income ofpartial specialists exceeds that of general practitioners by 29.6 per cent. Thecorresponding figures from Table 44 are 78.1, 39.0, and 28.2.19 The existence of regional or size of community differences was tested by ananalysis of ranks of the same sort as that used in the Appendix to Chapter 5to test the significance of regional and size of community differences in averageincome.
tTl
TA
BLE
44—
Arit
hmet
icM
ean
Inco
me
Stan
dard
ized
with
Res
pect
to S
ize
of C
omm
unity
,by
Typ
e of
Pra
ctic
e-'
Phy
sici
ans,
1929
—19
86o
Ave
rage
1929
'930
1931
1932
1933
1934
1935
1936
1929—36
'Standardized'
arith
met
ic m
ean
inco
me1
(do
llars
)G
ener
al p
ract
ition
ers
4,39
13,
850
3,32
82,
563
2,39
22,
589
2,68
13,
044
3,10
5P
artia
lspe
cial
ists
5,29
24,
432
3,98
42,
993
2,77
53,
169
3,35
63,
901
3,73
8C
ompl
ete
spec
ialis
ts8,
041
6,31
15,
130
3,78
734
933,
979
4,67
35,
421
5,10
4ce
.,
exce
ssof
Part
ial s
peda
lists
ove
r ge
nera
l pra
ctiti
oner
s20
.515
.119
.7i6
.8i6
.o22
425
.228
.220
42Z
Com
plet
e sp
ecia
lists
ove
r ge
nera
l pra
ctiti
oner
s83
.154
.147
.846
.053
.774
.378
.164
42C
ompl
ete
spec
ialis
ts o
ver
part
ial s
peci
alis
ts51
.942
.428
.826
.5.2
5.9
25.6
39.2
39.0
36.5
2
Wei
ghte
d av
erag
es o
f av
erag
es f
or s
ize
of c
omm
unity
cla
sses
.2C
ompu
ted
from
the
aver
age
inco
mes
in th
e fi
rst t
hree
line
s.Fo
r al
l yea
rs a
nd ty
pes
of p
ract
ice,
wei
ghts
are
the
tota
l num
ber
of p
hysi
cian
s in
eac
h si
ze o
f co
mm
unity
cla
ss in
193
1(s
ee0
Lel
and,
Dis
trib
utio
n of
Phy
sici
ans)
.
-4
276 PROFESSIONAL INCOMEever, many of the 'cells' contained few returns. Consequently,the ratios show erratic fluctuations that may have preventedsmall but real discrepancies from being detected.2°
Table 45 and Chart 24 show the relation between numberof years in practice and type of practice. The percentage ofgeneral practitioners declines at first, reaching a trough ofabout 28 per cent in the 13—17 years-in-practice class, and risesthereafter. The percentage of complete specialists, on theother hand, rises at first, reaching a peak of about 30 per centin the 18—22 years-in-practice class, and declines thereafter.The percentage of partial specialists fluctuates irregularlyaround 38 per cent. The striking feature of the chart is theexceedingly small percentage of specialists among physicianswho have been in practice more than 37 years. In part, thismerely reflects the relatively large percentage of the olderphysicians who practise in small communities in which spe-cialization is rare. But this is not the whole story. In corn-20 The frequency distributions of the ratios for individual size of communityand regional cells indicate the wide variability of the ratios. The frequencydistributions-ire given below. 1 stands for average income and the subscriptsc, p, and g, for complete specialists, partial specialists, and general practitionersrespectively.
NO. OF SIZE OF COMMUNITY AND REGIONAL CELLSFOR WHICH INDICATED RATIOS FALL BETWEEN
SPECIFIED VALUES
1 1 1VALUE OF RATIOS .72. ...2.
L9 iD
O—.5 3 5 1
6 7 11
1.0—1.5 11 i8 21
1.5—2.0 5 10 11
2.0—2.5 10 3 72.5—3.0 4 5 2
3.0—3.5 3 1
3.5—4.0 34.0—4.5 .. .. 1
4.5—5.0 1
Over5 .. 1
All values 46 46 54
The differences in the total number of cells arise from the absence in somecells of returns from one or more types of practitioner.
DETERMINANTS OF PROFESSIONAL INCOME 277
munities of each size, there is the same tendency toward adrastic decline in the percentage of specialists. The basic rea-son seems rather to be that mentioned earlier: the substantialimprovement in the standards of medical training since 1910which implemented the increasing need for specialists arisingfrom the rapid progress of medical science.
For 1936, we have computed average incomes by numberof years in practice and can therefore estimate the influenceof the concentration of complete specialists in the intermedi-
TABLE 45
Distribution by Type of Practice, by Years in Practice
Physicians, 1936% OF ALL PHYSICIANS WHO WERE
YEARS IN General Partial CompletePRACTICE practitioners specialists specialistsUnder 3 42.9 31.2 25.9
3— 7 474 354 17.28—12 33.8 40.4 25.8
13—17 28.5 44.3 27.318—22 31.2 38.8 29.923—27 37.1 34.5 28.328—32 38.5 39.8 21.6
33—37 40.7 41.8 17.538—42 35.0 g.643—47 56.9 39.3 3.748—52 56.7 43.353—57 76.2 23.8
In computing the percentages, returns for which number of years in practice wasunknown were excluded. The actual number of persons reporting in each classis shown in Table 34b, c, and d.
ate years-in-practice classes in which average incomes are rela-tively high. The accompanying table compares the percentagedifferences among average incomes standardized for differencesin the distribution of the various types of practitioner not onlyby size of community but also by years in practice2' with the21 The actual standardized averages are: complete specialists, $5,504, partialspecialists, $4,204, and general practitioners, $3486. These averages are basedonly on those size of community and years-in-practice classes containing re-turns from all three types of practitioner. They are weighted averages of theaverage incomes in such ceUs, the weights being the estimated number ofphysicians of all types in each cell.
278 PROFESSIONAL INCOMEpercentage differences previously presented for 1936. Thethird line of this table gives estimates of the percentage dif-ference between the incomes of physicians in communities
PERCENTAGE EXCESS IN 1936 AVERAGE INCOME OF
COMPLETE COMPLETE PARTIAL
SPECIALISTS SPECIALISTS SPEcIALISTS
OVER OVER OVER
GENERAL PARTIAL GENERAL
PRACTITIONERS SPECIALISTS PRACTITIONERS
Original averages 99.2 46.5
Averages standardized withrespect to distribution by:
Size of community 78.1 39.0 28.2
Size of community andno. of years in practice 57.9 30.9 20.6
of the same size and in practice the same number of years butdiffering in type of practice. Correction for the influence ofnumber of years in practice reduces the percentage differencesby about the same amount as correction for the influence ofsize of community. The final differences are much more mod-erate than the differences among the original averages.
Presumably, the differences that remain are largely attribu-table to differences in training and skill. Physicians whospecialize often get additional training; frequently they be-come specialists because they have been successful and haveattained good reputations as general practitioners. Hence,men who specialize would probably earn higher incomes asgeneral practitioners than those who remain general practi-tioners. The higher income of specialists is probably not atransitory phenomenon that will lead to or be eliminated bya rush to specialization; but rather a permanent concomitantof a segregation of practitioners by criteria related to theirchances of success.22
22 This interpretation is supported by the apparent temporal stability in theproportion of graduates who become or plan to become specialists suggestedby Weiskotten's study (see footnote 14).
The growing tendency in recent years to formalize the distinctions amongtypes of practitioner by establishing boards to 'certify' specialists in particularfields may introduce an additional element of rigidity.
280 PROFESSIONAL INCOME
TABLE 46
Percentage of Persons in Independent Practice Who are Membersof Firms and Number of Persons Covered
Certified Public Accountants, Lawyers, and Consulting Engineers
PROFESSION & SAMPLE 1929 1930 1931 1932 1933 1934 7933 1936
% Who are Members of FirmsCertified public accountants
'938 48.8 48.3 47.3 47.3'935 32.1 31.5 30.21957 35.9 30.7 32.5 85.3
Lawyers1955 23.8 23.0 22.71957 28.7 284 25.3 25.1 26.3 25.3
Consulting engineers44.2 44.8 48.5 43.5
Number of Persons CoveredCertified public accountants
1988Individual practitioners 373 593 409 426Firm members 590 609 6i 1 6371935Individual practitioners 792 842 883Firm members 623 647 635
1937Individual practitioners 843 525 535Firm members 346 895
Lawyers1935Individual practitioners 694 743 771Firm members 575 589 6061937Individual practitioners 504 566 6g6 676 733Firm members 220 239 249 258 285 303
Consulting engineers1938Individual practitioners 263 268 s6g 268Firm memb'ers 209 213 207 206
0 Number of persons covered by returns used, before weighting or adjusting. These num-bers, therefore, cannot be used to compute percentage of persons who are members of firms.
tice to small segments of the entire field. With these qualifica-tions, a classification by organization of practice is essentiallysimilar to a classification by type of practice. Like specialistsand for the same reasons, firm members are concentrated inthe larger communities, are seldom in the initial stages of their
DETERMINANTS OF PROFESSIONAL iNCOME 281
careers, and tend to receive higher average incomes than othermembers of the profession in the same community and of thesame age.
The percentage of all practitioners who are members offirms, as estimated from our samples, is given in Table 46 foreach year and profession.23 The different accountancy samplesyield widely divergent estimates of the percentage of firmmembers. According to the sample, slightly under 50 percent of all practitioners are firm members; according to theother two samples, only slightly over per cent. This dif-ference is much too large to be attributed to chance fluctua-tion; 24 and it cannot be interpreted as reflecting a markeddecline in the percentage of firm members, since for 1932 and
28 Information on the number of members in a firm was requested separately foreach year for which income information was obtained.24 Because of the weighting introduced to correct for the firm member bias, itis difficult to make a iogicaiiy valid test of the significance of the differencesamong the successive samples in the estimated proportion of firm members. Asan approximate test we computed x2 for the accompanying table.
NO. OF FIRM
MEMBERS
ACTUAL NO. OF WEIGHTED TO
SAMPLE AND YEAR INDIVIDUAL PRACrI- CORRECT FOR
FOR WHICH INCOME TIONERS REPORTINC FIRM MEMBER
WAS REPORTED INCOME BIAS TOTAL
l933samplefor 1932 426 382 8o8
1935 sample for 1934 883 1,266
1937 samplefor 1936 531 289 820
Total 1,840 1,054
The next to the last column gives our estimates of the number of firm memberswho would have replied if the questionnaire had requested information aboutthe individual recipient rather than the firm of which he is a member. Thelogical difficulty with using these estimates is that they are actually based onfewer separate returns, and hence are less accurate than the numbers alonemight suggest. Their use tends to accentuate the significance of the differences.However, since is 63.8, while a value of would be exceeded only oncein a thousand times, it seems clear that allowance for this difficulty would leavethe differences statistically significant.
282 PROFESSIONAL INCOME1929 we have estimates from two samples.25 We are forced toconclude that the difference is attributable to a bias in one ormore of our accountancy samples. In view of the marked agree-ment between the second sample, the largest of the three, andthe latest sample, we are disposed to set the proportion of allcertified public accountants who are members of firms at aboutone-third.
In all three professions the proportion of firm membersseems to have decreased somewhat from 1929 to 1932. In thetwo professions for which the data cover a longer period—lawand accountancy—this decrease continues until 1934. Fromthen to 1936 the proportion of firm members increases. Thesechanges are all exceedingly small, but their consistency is areason for believing that they reflect the facts accurately.Moreover, it seems reasonable that firms are more likely todisintegrate and less likely to be formed when economic con-ditions are becoming worse than when they are improving.
Table 47 reveals that the average incomes of firm membersare considerably higher than those of individual practitioners.The differentials vary considerably from sample to sample.This time the second sample of accountants disagrees with theother two; the first and third samples suggest that the averageincome of firm members is about 25 or 30 per cent larger thanthe average income of individual practitioners; the secondsample, that it is almost 6o per cent larger. For lawyers, the
25 The lapse of time between the dates at which the two samples were chosenmight well have led to an underestimate of the proportion of firm members inthe overlapping year from the later sample and an overestimate from the earliersample, for reasons discussed in Chapter 2. The proportion of finn memberswould presumably be relatively high among individuals retiring from practicebetween say, the first sample was chosen—and 1935—when thesecond sample was chosen—and relatively low among individuals inadvertentlyexduded from the 1933 samplinglist because of recent entry into the profession.But it is hardly credible that this factor could account for so large a differenceas that between the 1932 estimates from the 1933 and 1935 samples since onlytwo years separated the dates at which these samples were chosen. It is some-what more reasonable, though still exceedingly doubtful, that the differencebetween the t 929 estimates from the 1933 and 1937 samples can be explainedin this way.
DETERMINANTS OF PROFESSIONAL INCOME 283
TABLE 47
Arithmetic Mean by Organization of
Certified Public Accountants, Lawyers, and Consulting Engineers
PROFESSION &&tMPLE 1929 1930 '93' 1932 '933 1934 1935 1936
Arithmetic Mean Income (dollars)Certified public accountants
1933Individual practitioners 6,941 6s8g 4,313Firm members 8.962 8,414 5,294'935Individual practitioners 3,275 5,664Firm members 5,645 5,6821937Individual practitioners 5,469 3,81', 4,248Firm members 4,803 4,926 5,120
Lawyers1955Individual practitioners 2,883 2,515 2,656Firm members 5,511 5,048 5,2601957Individual practitioners 5,762 3,205 *,8s6 2,848 *,gg6 5,1.89Fix-rn members 15,973 io,6&3 9,859 9,706
Consulting engineers1935IndividuaL practitioners 8,555 7,472 4,945 2.483Firm members i6,oig 13,264 7,114 3,940
% by which Arithmetic Mean Income of Firm Members ExceedsArjthnietjc Mean Income of Individual Practitioners
Certified public accountants1933 29.1 81.6 22.71935 59.2 59.2 55.11957 19.8 32.7 29.1 *0.5
Lawyers1955 91.2 100.7 98.01957 142.5 231.1 248.9 240.8 228.6 2497
Consulting engineers1935 87.7 77.5 43.9 58.7
1935 sample suggests a difference of about ioo per cent; the1937 sample, if we ignore the comparison for 1929, a differ-ence of almost 250 per cent.26 The one sample for consultingengineers suggests a difference between 40 and 90 per cent.While there is some overlapping among the several profes-sions it seems clear that the difference between firm membersand individuals is greatest for lawyers and least for accountants.
According to the measures in Tables 48 and 49, absolute
28 However, if the one extreme questionnaire repeatedly mentioned is elimi-nated the difference is reduced to about 125 per cent.
t0 00
TA
BLE
48
Med
ian
and
Qua
rtile
Inc
omes
, by
Org
aniz
atio
n of
Pra
ctic
e
Cer
tifie
d Pu
blic
Acc
ount
ants
, Law
yers
, and
Con
sulti
ng E
ngin
eers
1933
SA
MP
LES
1935
SA
MP
LES
1937
SA
MP
LES
1929
1930
1931
1932
1932
1933
1934
1934
1935
1936
(dol
lars
)
Thi
rd Q
uart
ileC
ertif
ied
publ
ic a
ccou
ntan
tsA
fl9,
308
8,56
07,
326
5,99
35,
029
4,72
45,
232
4,96
75,
073
5,68
7
Indi
vidu
al p
ract
ition
ers
8,56
57,
670
6,25
25,
486
4,32
53,
908
4,54
14,
494
4,75
45,
384
Firm
rae
mbe
rsg,
g68
9,37
28,
268
6,50
96,
565
6,03
16,
502
5,92
65,
989
6,29
5
Law
yers
All
4,33
93,
620
3,93
60
Indi
vidu
al p
ract
ition
ers
3,33
43,
010
3,19
2
Firm
mem
bers
7,00
86,
241
7,10
8
Con
sulti
ng e
ngin
eers
All
14,8
0511
,721
8,63
14,
785
0In
divi
dual
pra
ctiti
oner
s10
,650
9,38
46,
411
3,95
5Fi
rm m
embe
rs18
,800
14,8
9612
,500
5,73
5
Med
ian
'-'
Cer
tifie
dpu
blic
acc
ount
ants
Z
All
6,ii6
5,64
74,
780
4,01
73,
336
3,12
93,
515
3,35
83,
460
3,96
3
Indi
vidu
al p
ract
ition
ers
5,18
65,
000
4,15
73,
509
2,88
42,
763
3,12
23,
099
3,29
03,
660
Firm
mem
bers
6,72
76,
322
5,41
14,
452
4,62
44,
288
4,87
64,
292
4,06
94,
514
Law
yers
All
2,21
8t,g
o62,
028
Indi
vidu
al p
ract
ition
ers
i,goo
1,58
11,
747
Firm
mem
bers
4,00
43,
275
3,74
2
Con
sulti
ngen
gine
ers
All
7,94
36,
016
4,04
12,
178
ZIn
divi
dual
prac
titio
ners
6414
5,12
53,
190
1,52
4Fi
nn m
embe
rs9,
545
6,84
16,
125
3,30
0
Firs
tQua
rtile
Cer
tifie
d pu
blic
acc
ount
ants
oA
ll4,
099
3,88
33,
217
2,54
12,
235
2,12
72,
356
2,29
62,
479
2,68
4In
divi
dual
prac
titio
ners
3,62
93,
297
2,78
92,
294
1,97
51,
936
2,07
42,
143
2,30
42,
532
Firm
mem
bers
4,39
44,
380
3,77
62,
985
3,16
22,
886
3,23
62,
747
2,86
13,
111
Law
yers
Mi
1,14
098
296
7In
divi
dual
prac
titio
ners
840
854
Firm
mem
bers
2,14
91,
778
0C
onsu
lting
eng
inee
liZ
All
3,57
02,
719
1,45
6In
divi
dual
pra
ctiti
oner
s2,
769
2,00
057
00
Firm
mem
bers
4,86
74,
343
2653
575
z c,J 0 10 0O C;'
TA
BL
E 4
900
Mea
sure
s of
Var
iabi
lity
of I
ncom
e, b
y O
rgan
izat
ion
of P
ract
ice
Cer
tifie
d Pu
blic
Acc
ount
ants
, Law
yers
, and
Con
sulti
ng E
ngin
eers
1933
SA
MPL
ES
1935
SA
MPL
ES
1937
SA
MPL
ES
1929
1930
'93'
1932
1932
1933
1934
1934
1935
1936
Certified
publ
ic a
ccou
ntan
tsln
terq
uart
ile D
iffe
renc
e (d
olla
rs)
All
5,20
94,
677
4,10
93,
452
2,79
42,
597
2,87
62,
671
2,59
43,
003
Indi
vidu
al p
ract
ition
ers
4,93
64,
373
3,46
33,
192
2,35
01,
972
2,46
72,
351
2,45
02,
852
Firm
mem
bers
5,57
44,
992
4,49
23,
524
3,40
33,
145
3,26
63,
179
3,12
83,
184
Law
yers
All
3,19
92,
638
2,96
9In
divi
dual
pra
ctiti
oner
s2,
364
2,17
02,
338
Firm
mem
bers
4,85
94,
463
5,21
6C
onsu
lting
eng
inee
rsA
ll11
,235
9,00
27,
175
4,75
2In
divi
dual
prac
titio
ners
7,88
17,
384
5,84
13,
955
Firm
mem
bers
13,9
3310
,553
9,84
75,
160
0
Cer
tifie
d pu
blic
acc
ount
ants
Stan
dard
Dev
iatio
n (d
olla
rs)
All
6,72
36,
410
5,15
23,
708
3,56
83,
360
3,48
33,
072
3,33
43,
240
Indi
vidu
al p
ract
ition
ers
5,81
04,
856
4,40
73,
266
3,21
12,
937
3,01
92,
866
2,92
53"
°30
Firm
mem
bers
7481
7,53
45,
819
4,07
53,
881
3,81
54,
023
3,35
23,
956
3,41
0 Z
Law
yers
All
4,36
94,
368
4,16
4In
divi
dual
pra
ctiti
oner
s3,
921
3,85
93,
713
Firm
mem
bers
5,09
95,
306
4,92
1Z
Consulting
engi
neer
sC
,
All
14,5
80i6
,669
9,01
06,
462
0In
divi
dual
pra
ctiti
oner
s10
,363
10,7
249,
391
6,13
9Fi
rm m
embe
rs17
,580
21,6
158,
455
6,83
5
Rel
ativ
e In
terq
uart
ile D
iffe
renc
e.C
ertif
ied
publ
ic a
ccou
ntan
tsA
ll.8
52.8
28.8
6o.8
38.8
30.8
i8.7
95.7
50In
divi
dual
pra
ctiti
oner
s.9
52.8
74.8
32.9
10.8
15.7
14.7
90.7
59.7
45.7
79Fi
rm m
embe
rs.8
29.7
90.8
30.7
92.7
36.7
33.6
70.7
41.7
69.7
05La
wye
rsA
ll1.
384
1464
Indi
vidu
al p
ract
ition
ers
1.33
8Fi
rm m
embe
rs1.
214
1.36
31.
394
ZC
onsu
lting
engi
neer
sA
ll14
141.
496
1.77
62.
182
Indi
vidu
al p
ract
ition
ers
1.22
81.
850
2.59
6Fi
rm m
embe
rs14
601.
543
i.6o8
1.56
4
Rat
io o
f Q
uart
iles
Cer
tifie
d pu
blic
acc
ount
ants
oA
ll2.
271
2.20
42.
279
2.35
72.
250
2.22
12.
221
2.16
32.
046
2.11
9In
divi
dual
prac
titio
ners
2.36
02.
326
2.24
22.
591
2.19
02.
019
2.18
92.
097
24)6
32.
126
Firm
mem
bers
2.26
92.
140
2.19
02.
181
2.07
62.
090
2.00
92.
157
2.09
32.
023
Law
yers
All
4.07
00
Indi
vidu
al p
ract
ition
ers
5457
3.75
8Fi
rm m
embe
rs3.
261
3.51
03.
757
Con
sulti
ng e
ngin
eers
All
4.14
74.
311
5.92
814
5.00
0In
divi
dual
pra
ctiti
oner
s5.
846
4.69
211
.247
00Fi
rm m
embe
rs3.
863
3430
4.71
29.
974
0
Coe
ffic
ient
of
Var
iatio
n•C
ertif
ied
publ
ic a
cCou
ntan
tsA
ll.8
43.8
43.7
68.8
40.8
54.7
61.7
84.7
00In
divi
dual
pra
ctiti
oner
s.8
37.7
72.8
54.7
57.8
93.8
83.8
i 1.7
8o.7
50.7
1700
Firm
mem
bers
.854
.883
.765
.6gi
.727
.698
.Sgi
.793
.659
Law
yers
All
Indi
vidu
al p
ract
ition
ers
Firm
mem
bers
Con
sulti
ng e
ngin
eers
All
Indi
vidu
al p
ract
ition
ers
Firm
mem
bers
• U
nder
the
assu
mpt
ions
out
lined
in th
e fo
otno
te to
Tab
lew
e ca
n es
timat
e ro
ughl
y th
e ef
fect
, on
the
coef
fici
ent o
f va
ria-
tion,
of
the
dow
nwar
d bi
as a
risi
ng f
rom
div
idin
g th
e in
com
e of
a fi
rm e
qual
ly a
mon
g its
mem
bers
. If
all f
irm
s co
ntai
ned
two
mem
bers
, and
if th
ere
wer
e no
cor
rela
tion
betw
een
the
inco
mes
of m
embe
rs o
f th
e sa
me
firm
, the
'tru
e' s
tand
ard
devi
atio
n fo
r
1933
SA
MPL
ES
ASS
UM
ED
CO
RR
EL
AT
ION
1929
1930
1931
Coe
ffic
ient
of
Var
iatio
n(c
onL
)
1.23
61.
390
1.26
61,
340
1.50
81.
377
.930
1.04
0.9
29
1937
SA
MPL
ES
1934
1935
1936
CI)
0l
Accountants
.797
.839
.797
.915
.760
.977
1.02
8.9
87.9
7711
2193
2t.4
Law
yers
1.07
31.
200
1.07
2L
315
1471
1.31
4
Eng
inee
rs
TA
BL
E 4
9 (c
ond.
)1935 SAMPLES
1932
1933
1934
1933
SA
MPL
ES
1929
1930
1931
1932
1.231
1.64
71.
548
2.12
41.
214
1435
1.90
02.
472
1.12
01.
635
1.23
91.
759
1937
SA
MPL
ES
1934
1935
10 00 0019
36
firm
mem
bers
, and
con
sequ
ently
the
'true
' coe
fild
ent o
f va
ria-
tion,
wou
ld te
nd to
be
141
(squ
are
root
of
2) ti
mes
the
mea
sure
sas
we
com
pute
them
. If
the
corr
elat
ion
wer
e .5
, the
cor
rect
ion
fact
or w
ould
be
1.15
. App
lyin
g th
ese
fact
ors
give
s th
e ac
com
-pa
nyin
g re
sults
for
the
coef
fici
ents
of
vari
atio
n fo
r fi
rm m
embe
rs.
.962
L01
91.
179
1.24
9
1935
SA
MPL
ES
1932
1932
1933
1934
.5 0 -5 0 .5 0
.965
.883
1.18
2i.0
8a
0
1.29
2i.8
8814
302.
030
1.58
42.
313
1.75
224
87
z 0
DETERMINANTS OF PROFESSIONAL INCOME 289
variability of income is larger for firm members than for indi-vidual practitioners but relative variability is smaller. Themeasures for firm members are, however, much affected by thedownward bias discussed in Chapter 4, Section ib. Its elimina-
TABLE 50
Percentage of Persons in Independent Practice Who Are Membersof Firms, by Size of Community and by Region
Certified Public Accountants, Lawyers, and Consulting Engineers
% WHO ARE MEMBERS OF FIRMS1Consulting
Certified public accountants Lawyers engineersSize of community 1932 1934 1936 1934 1936 1932
& over 53.0 30.9 33.5 25.1 33.9500,000—1,500,000 39.0 33.4 34.9 4.6 22.3 42.1250,000— 500,000 55.8 37.1 47.0 23.4 31.0 57.9100,000— 250,000 46.3 35.0 30.5 30.6 i8.g 29.7
25,000— 100,000 51.1 25.4 40.2 30.0 34.0 45.710,000— 25,000 ) 27.1 23.0
2,500— 10,000 11.3 12.2 92.1 23.9 20.3Under 2,500 .) 17.8 11.0
RegionNew England 42.7 33.6 34.3 22.0 iS.6 46.5Middle Atlantic 49.6 30.4 34,3 20.8 18.8 55.6E. N. Central 53.3 37.0 41.0 21.0W. N. Central 48.8 42.0 54.7 27.3 18.4 40.0S. Atlantic 54.2 90.3 34.1 21.5 95.9
2 0E. S. Central 45.4 32.1 31.5 33.2 9.7 jW. S. Central 26.2 55.1 30.5 20.8 40.0Mountain 56.9 22.1 35.5 19.9 19.5 40.0Pacific 36.7 22.3 24.0 15.5 39.9 31.4
U. S.1 47.3 30.2 35.3 22.7 25.5 45.5
1 Percentages are based on numbers reporting for the last year covered by each sample.2 Includes a few returns for which size of community or region was unknown.
tion would increase the difference in absolute variability, butmight well erase or even reverse the observed difference inrelative variability.27
Organization into firms is least frequent in small communi-ties (Table 50). There is little difference among the profes-
27 Rough estimates of coefficients of variation corrected for this bias are givenin the footnote to Table 4g,. For accountants the coefficients of variation forindividual practitioners are below both sets of corrected coefficients for firmmembers for the first and third and between the two sets for thesecond sample. For the coefficients of variation for individual practi-tioners exceed both sets of corrected coefficients for firm members. For engineersthe coefficients for individual practitioners are below both sets of corrected
290 PROFESSIONAL INCOMEsions in this respect. Except for the single erratic value fromthe 1935 legal sample, firm members are uniformly leastnumerous in communities with populations of less than25,000; the variation among the other size of communityclasses in the percentage of firm members seems to be random.The regional differences in the percentage of firm members,while larger than can be accounted for by chance, are ingeneral less consistent from profession to profession than thedifferences among the size of community classes.28
In the smallest communities, where firms are few, the ad-vantage of organization into firms is apparently least; indeed,according to three of the six samples for which data are givenin Table 51, firm members in the smallest communities actu-
coefficients for firm members for two years, between them for one, and abovethem for one. This evidence suggests that, in accountancy, relative variabilityof income is greater for firm members than for individual practitioners; in law,the situation is reversed, and in engineering, relative variability is about thesame for individual practitioners and firm members. The roughness of ourestimates of the corrected coefficients of variation, the margin of error attach-ing to the original measures, and the neglect of other measures of variabilitymean that these condusions must be considered exceedingly tentative.28 To test the significance of the differences between the distributions of individ-ual practitioners and firm members by size of community and region we usedthe test. In each case x2 was computed from a table giving in one columnthe number of individual practitioners in the sample in each size of community(or region), and in a second column, the number of firm members.
accountants we used the actual number of individual practitioners, butthe number of firm members weighted tO correct for the firm member bias.For the 1935 legal sample, we used the number of firm members weighted tocorrect for the firm member bias but we adjusted neither the number ofindividuals nor the number of firm members for the size of community bias.The original number of individuals and of firm members was used for the 1937legal sample, no adjustment being made for either the size of community biasor the nonrandomness of the sample by states. The use of figures for lawyersnot adjusted for the size of community bias and the nonrandomness of the1937 sample by states is partly justified by the 'null' hypothesis being tested—namely, that the proportion of firm members is the same in all regions or in allsize of community classes. In any event, this procedure is almost unavoidableif any attempt is to be made to interpret the results in probability terms. Thesample of consulting engineers is the only one that raises no problems, sinceit is subject to no biases for which we have attempted to adjust. The values ofxa are summarized in the accompanying table. For reasons given in footnote 24,the procedure used led to overestimates of the significance of the differences,
DETERMINANTS OF PROFESSIONAL INCOME 291
ally have lower average income than individual practitioners.There is some indication that the advantage of organizationinto firms increases consistently with size of community. Con-sulting engineers seem to be an exception to this statement;but no reliance can be placed on the figure in Table 51 forconsulting engineers in communities under 25,000. The firm
since in all cases, the number of firm members is greater than the number ofseparate returns for firm members.
TEST OF SIGNIFICANCE OF DIFFERENCE BETWEEN
DISTRIBUTIONS OF FIRM MEMBERS AND OFINDIVIDUAL PRACTITIONERS BY
SIZEOFCOMMUNITY REGIONNo. of No. of
degrees of degrees offreedom freedom x2
Certified public accountantssample for 1932 5 8 14.9
1935 sample [or 1934 5 8 21.9j1937 sample for 1936 5 14.9* 8
All samples 8Lawyers
1935 sample for 1934 7 81937 sample for 1936 7 8
Both samples 7 8 14.9Consulting engineers
1933 sample for 1932 27.4 7 iB.8f
Greater than the value that would be exceeded by chance once in twentytimes.f Greater than the value that would be exceeded by chance once in a hundredLimes.
Greater than the value that would be exceeded by chance once in a thousandtimes.
For 5 degrees of freedom these values are i 1.070, 15.086, and 20.517 respec-tively; for 7 degrees of freedom, 14.067, 18.475, 24.322; for 8 degrees of freedom,15.507, 20.090, 26.125.
On this showing, the significance of the difference between the distributionof firm members and individuals by size of community is indisputable, eventhough large allowance is made for deficiencies in the tests. The differencesbetween the regional distributions are less marked, though it seems fairly certainthat on the whole they are larger than could be expected from chance alone.In each comparison the value of x2 is greater for the size of community dis-tributions than for the regional, although more degrees of freedom are avail-able for the latter.
S
292 PROFESSIONAL INCOMEmember average used in deriving this figure is based on onlyone firm for the two years that are responsible for its relativelyhigh value.29
As already suggested, it seems reasonable to interpret theseresults in terms of the clientele to whom services are rendered.
TABLE 51
Difference between Arithmetic Mean Income of Firm Membersand of Individual Practitioners
Certified Public Accountants, Lawyers, and Consulting Engineers
% BY WHICH ARITHMETIC MEAN INCOME OF FIRM MEMBERS EXCEEDSARITHMETIC MEAN INCOME OF INDIVIDUAL PRACTITIONERS
ConsultingCertified public accountants Lawyers engineers
Size of community 1929—32 1932—3 4 1934—36 1932—34 1934—36 1929-32
I,500,000&over 40.8 88.5 56.0 197.9 456.1 82.8500,000-1,5o0,000 17.6 30.2 59.5 336.1 195.9250,000— 500,000 40.5 70.1 33.5 99.1 168.3 —18.4100,000— 250,000 ii.8 29.3 2.4 65.4 54.6 —1.4*5,000- 100,000 —0.4 13.9 —9.1 90.4 128.8 —38.810,000- 25.000 11.2 109.52,500- 10,000 —5.9 —1.5 18.9 63.7 47.8 so8.8
Under .1 —28.0 J
RegionNew England 51.8 48.2 35.6 296.8 232.9 6i.8Middle Atlantic 51.5 79.0 85.2 117.1 209.4 95.6E. N. Central 54.8 58.8 10.2 115.7 —25.5W. N. Central —14.4 43.4 54.7 46.5 24.2 64.9S. Atlantic 55.9 17.4 —2.1 66.2 185.6 1E. S. Central 47.6 78.3 64.5 83.4 .3
W. S. Central 15.2 74.0 35.2 94.0 —28.2 —52.9MountaIn 4.0 66.g 87.6 31.5 —64.8Pacific i6.8 42.4 9.3 i6g.i 9.0
U. S.' 29.8 57.8 27.1 96.4 259.9 72.1
1 For all professions, percentage difference is computed from simple unweighted averages ofthe annual averages.I Based on averages that include a few returns for which size of community or region wasunknown.
In small communities much more than in large, the clientele islikely to be composed of small businesses and individual con-sumers; and the 'unit' of service is likely to be small and per-sonal. Organization into firms is less advantageous and 'lessfrequent.
The difference between the average incomes of firm mem-
29 The percentage differences for the individual years are of interest in this1929, 474.5; 1930, 167.2; 1931, —342.5; 1932, _i8i.g.
DETERMINANTS OF PROFESSIONAL INCOME 293
bers and individual practitioners in communities of the samesize is in general less than the difference between the country-wide averages. (Compare Tables 47 and 52.) For accountants,correction for the influence of size of community reduces thedifference between the incomes of firm members and mdi-
TABLE 52
Difference between Arithmetic Mean Incomes of Firm Membersand of Individual Practitioners, Based on Averages Standardizedwith Respect to Size of Community
Certified Public Accountants, Lawyers, and Consulting Engineers
% BY WHICH STANDARDIZED ARITHMETIC MEAN INCOMEOF FIRM MEMBERS EXCEEDS STANDARDIZED ARITHMETIC
MEAN INCOME OF INDIVIDUAL PRACTITIONERS1
PROFESSION & SAMPLE 1929 1930 1931 1932 1933 1934 1935 1936Certified public accountants1
1933 25.1 32.7 24.6 15.51935 53.3 54.3 51.21937 30.6 27.9 19.6
Lawyers'1935 102.0 108.9 105.21937 142.1 169.0
Consulting engineers'1933 944 54.6 10.1 39.3
1 standardized averages from which the percentage differences are computedare weighted averages of the averages for the size of community classes. Thesame weights are used for both individual practitioners and firm members.'The weights used in computing the standardized averages are, for each size ofcommunity class, the total number of accountants, adjusted for the firm memberbias, in both the 1935 and samples who reported incomes.'The weights used in computing the standardized averages are, for each size ofcommunity dass, the total number of lawyers, adjusted for the size of com-munity and firm member biases, in the 1935 sample who reported 1934 incomes.'The weights used in computing the standardized averages are, for each size ofcommunity class, the total number of engineers reporting 1932 incomes.
vidual practitioners only slightly: in 1936, the year for whichthe reduction is least, from 20.5 to ig.6, but in the 1932 aver-ages from the 1933 sample, the year and sample for which thereduction is greatest, from 22.7 to 15.5, i.e., by almost a third.The correction for the influence of size of community actuallyincreases the differential shown by the first legal sample. Thisanomalous result is entirely accounted for by one size of corn-
294 PROFESSIONAL INCOMEmunity class—500,000— 1,500,000. The percentage of firm mem-bers in this class in the 1935 sample is exceedingly 1OW—4.6per cent in 1 934—but the average income of the firm membersreporting was 323 per cent higher than the average income ofthe individual practitioners.3° The '937 legal sample yields avery different result; correction for the influence of size ofcommunity reduces the differential by a third. As usual, theengineering sample yields erratic results: the differential
TABLE 53
Distribution by Organization of Practiceand Year of Admission to Bar
New York County Lawyers,FIRM MEMBERS
NO. IN ARE OF ALL
YEAR OF SAMPLE IN NUMBER OF LAWYERS IN
ADMISSION INDEPENDENT INDIVIDUAL FIRMTO BAR PRACTICE MEMBERS PRACTICE
1930—1934 730 583 147 20.11924—1929 1,023 747 276 27.01918—1923 433 275 158 36.5
1911—1917 577 250 147 39.0
1900—1910 508 339 169 33.3
Before 1900 257 152 105 40.9
All years 3,328 2,326 1,002 30.1
Survey of the Legal Profession in New York County, p. 28.
in 1929 is increased from 88 to 94 per cent; the differential in1931 is reduced from 44 to 10 per cent.
We have evidence on the relation between organization ofpractice and number of years in practice only for lawyers. Astudy of New York County lawyers indicates that the percent-age of firm members at first increases sharply with number ofyears in practice and then remains fairly constant (Table 53).Studies of young lawyers in California and Wisconsin confirmthe initial sharp rise.3' As we saw in Section 2, the average in-
30 The number of firm members in that class in 1934, weighted for the firmmember bias, was 5.31 See summary of restilts for California and Wisconsin lawyers in Economics ofthe Legal Profession, p. 48.
DETERMINANTS OF PROFESSIONAL INCOME 295
come of lawyers in the years-in-practice groups in which firmmembers are most numerous is considerably higher than inthe years-in-practice groups in which firm members are fewest.Consequently, the difference between the average incomes ofindividual practitioners and firm members in practice thesame number of years would be considerably smaller than thedifference between the countrywide averages; or between theaverages standardized with respect to size of community dis-tribution. Unfortunately, the available data are too meagrefor even rough estimates of the quantitative effect of this factor.
c Salaried and independent practiceAs noted above, independent practice is likely to dominate pro-fessions that serve primarily individuals; salaried practice,professions that serve primarily business enterprises and pub-lic agencies. Accordingly, independent practice is dominantamong physicians and dentists and the numerically smallgroups of authors, composers, etc. In contrast to these are ac-countants, engineers, chemists, metallurgists, etc., groups inwhich independent practice is relatively rare. Certified publicaccountants may seem an exception, since they are predomi-nantly individual practitioners. However, as previously noted,they belong to the same professional group as the large numberof accountants and auditors employed by business firms andgovernment. In any profession that is largely salaried, somemen, either because of superior skill or for other reasons, findit advantageous to practise without an attachment to a businessor governmental organization. Consulting engineers, econ-omists, chemists, etc., are other examples of such auxiliarygroups of independent practitioners.
Law individuals, business enterprises, and govern-mental agencies, and in consequence occupies an intermediateposition between these two groups. Initially dominated by in-dependent practice, it has become increasingly salaried aslarge business enterprises have grown in importance and thegovernment's role as an employer has expanded.
296 PROFESSIONAL INCOMEThe independent practitioner must make a larger capital
investment in equipment, expense of building up a practice,etc., than the salaried man, and must assume a greater risk.These factors that would tend to make the average income ofthe independent practitioner larger than the average incomeof the salaried man 82 are common to all professions. In addi-tion, in medicine and dentistry, and to a considerable extent,law, salaried employment is ordinarily a step toward inde-pendent practice. Salaried employees are therefore likely tobe younger than independent practitioners and to include alarger proportion of men who have not yet reached their peakearnings. The salaried employee of a physician, dentist, orlawyer is unlikely to receive more than his employer, thoughof course he may receive more than other independent practi-tioners. Since prestige plays so large a role in attracting custom,the salaried man will find it advantageous to enter independ-ent practice as he gains experience and acquires a reputationon his own account.
In accountancy and engineering, most men are salariedemployees throughout their professional career. The inde.pendent practitioners are an auxiliary group. The businessenterprises and governmental agencies that are the majorconsumers of the services of these professions are likely to pur..chase from independent practitioners solely highly specializedservices: services that their own professional employees cannotrender and that are required in such small amounts that it isnot profitable to employ additional full-time employees; orservices that they would prefer to have performed by outsideagencies whose findings will be respected as objective and im-partial (e.g., an independent audit). Independent practitionersin these professions are likely to be recruited from men whohave done particularly well as salaried employees and have be-
82 Independent practice may, of course, have greater appeal for other reasonsthan the expectation of larger pecuniary returns. But it may be hazarded thatthe 'net advantages' are not so dearly or generally on the side of independentpractice as to affect the general tenor of this analysis.
DETERMINANTS OF PROFESSIONAL INCOME 297
come fairly well known; men who would earn relatively highincomes as salaried employees.33
The relative status of the salaried and independent groupsmay be reversed during cyclical depressions when large num-bers of previously salaried individuals may become unem-ployed and enter independent practice because they find itimpossible to obtain salaried employment. Such a conditionis unlikely to become chronic, however, unless there is a defi-nitely inferior group of professionally trained persons who canmanage to stay in private practice but would not be hired byemployers, or unless competition plays little or no role in thepricing of professional services; for otherwise competition,slow and halting though its workings may be, will tend to drivedown salaries until the situation again approaches that out-lined above.
The incomes of independent practitioners are likely to benot only higher but also more variable than those of salariedemployees. Independent practitioners are likely to be a hetero-geneous group, including at the one extreme, some men whoare in independent practice as a temporary expedient becausethey cannot get a salaried post, at the other extreme, men whorender highly specialized services for which no one consumercan provide an adequate outlet. Salaried employees are or-dinarily a more homogeneous group: a man is not likely to beemployed at all unless he is worth the usual 'starting' salary;and while in time he may have a fairly responsible and well-paid position, he is unlikely to become a 'top' executive unlesshe subordinates professional work to general managerial ac-tivity. Two other factors, touched on in Chapter 4, Sectionalso make for greater variability of income from independentpractice. In the first place, the incomes of independent prac-titioners include an element of entrepreneurial return, a typi-
830f course, these considerations apply only to men engaged in renderingservices strictly related to their specialty; they do not apply to men trained,for example, as lawyers or accountants, who become corporate executives oroccupy other posts in which they perform tasks largely unconnected with theirparticular profession.
298 PROFESSIONAL INCOMEcally variable element greatly subject to 'random' influences.In the second place, 'nonrational' factors, which make for widevariation in consumers' judgments of the quality of the serv-ices rendered by different men, affect independent practition-ers more than salaried employees since the former are morelikely to sell services to individuals and in small quantities.
The National Resources Committee estimates of the dis-tribution of income by size, cited in Chapter 3, tend to supportthese statements. The arithmetic mean income of independ-ent professional families in 1935—36 was more than twice thatof salaried professional families, and the variability of incomewas greater among independent than among salaried families(Table 8 and Chart However, these data are for all pro-fessions combined, whereas the considerations just presentedapply to each profession separately. The larger and more vari-able incomes of independent professional families may merelymean that independent practitioners are concentrated in thoseprofessions in which incomes are largest and most variable.They do not necessarily mean that independent practitionersreceive larger and more variable incomes than salaried men inthe same profession.
The estimates summarized in Table 54 bear more directlyon the relative income of salaried and independent practi.tioners in the same profession, but are based on much smallersamples. They indicate that salaried employees tend to havelower arithmetic mean incomes than their independent breth-ren but may have higher median incomes. This difference re-flects, of course, a greater skewness in the distribution ofincome from independent practice. In general, the data on
the estimates in Table 54 are based show greater vari-ability of income from independent practice than from salariedemployment. The available data, though hardly adequate toestablish the conclusions suggested by the considerations out-lined above, are entirely consistent with them.
DETERMINANTS OF PROFESSIONAL INCOME 299
TABLE 54
Arithmetic Mean and Median Incomes of Salaried Employeesand Independent Practitioners
Physicians and Lawyers: Selected Studies
YEAR
TO WHICH ARITH.
ESTIMATES NO.OF MEAN MEDIAN
RELATE RETURNS INCOME INCOME
Physicians (dollars)
American Medical Association Study 1 1928
Salaried 853 5,428 4,718Independent 5,475 6,499 4,938
Committee on the Costs of Medical Care,'accepted' estimate 1929
Salaried 6 4,524 4,213Independent . . 5,467 3,705
California Medical-Economic SurveyFull-time salaried' . . . 3345 3,000Full- & part-time salaried . e 3,674 3,3wAll physicians 2,737 3,572 2,700
Law yers
New York County Study 6 1933
Employed in law offices on 'salary basis' 320 4,316 3,400All employed in law offices 558 4,011 2,885Independent 2,667 6,664 3,210
Incomes of Physicians, p. 1o5.a Ibid., p. 20. The 'accepted' estimate is based on the American Medical Associa-tion study and data for special groups of salaried physicians.3Catifornia Medical-Economic Survey, pp. 90 and 94.'Estimates of 1934 salary.
Survey of the Legal Profession in New York County, pp. i8, 34. This reportdoes not contain arithmetic averages for any group, or medians for salariedemployees. We computed them from frequency distributions. The lawyers em-ployed on other than a 'salary basis' presumably receive commissions or per.quisi tes. The averages are for total professional income, not only income fromsalaried employment.
not available.