This PDF is a selection from an out-of-print volume from the NationalBureau of Economic Research
Volume Title: The Personal Distribution of Income and Wealth
Volume Author/Editor: James D. Smith, ed.
Volume Publisher: NBER
Volume ISBN: 0-870-14268-2
Volume URL: http://www.nber.org/books/smit75-1
Publication Date: 1975
Chapter Title: Individual Taxes and the Distribution of Income
Chapter Author: Benjamin A. Okner
Chapter URL: http://www.nber.org/chapters/c3751
Chapter pages in book: (p. 45 - 74)
CHAPTER 3
Individual Taxes and theDistribution of Income
Benjamin A. OknerThe Brookings Institution
Most economists agree that the most appropriate way tomeasure the burden or incidence of taxes is in terms of their effecton the distribution of income. There is so little dispute, in fact,that incidence is usually defined simply as the effect of taxes onthe distribution of income available for private use.
In order to determine empirically the burden of any tax, someassumption must be made about whose real income it reduces andabout the amount of before-tax income that would have beenreceived in its absence. Since taxes affect the total level ofeconomic output as well as the distribution of real income throughtheir impact on both factor prices (the sources of income) andcommodity prices (the uses of income), a full-fledged study of theoverall burden of taxes is exceedingly complex.
In this paper, I am concerned only with the distributionaleffects of the personal income and employment taxes. Since thesetwo levies amounted to more than 50 percent of total governmenttax receipts in 1966,1 their incidence is of considerable impor-tance for any overall study of tax burdens. In addition, there is
The views presented are those of the author and not necessarily those ofthe officers, trustees, or other staff members of The Brookings Institution.All computer operations described in the paper were performed at theBrookings Social Science Computation Center and the programming was doneby Andrew D. Pike, whose efforts are gratefully acknowledged. The workdescribed is part of a research program supported by a grant from the U.S.Office of Economic Opportunity.
1 Total government tax receipts in the study are equal to federal, state,and local government receipts as measured in the National Income Accounts,adjusted to exclude nontaxes, intergovernmental grants-in-aid, and socialinsurance contributions for civilian government retirement funds and vet-erans' life insurance.
45
46 Benjamin A. Okner
little disagreement as to how these taxes should be allocatedamong persons.2
Once one has decided how a given tax should be allocatedamong individuals in the population, its incidence—or impact onthe distribution of income—might be presumed to be a ratherstraightforward calculation. However, this is not necessarily thecase, because there are a number of ways in which income mightbe defined. Although we prefer a particular income definition, atthe beginning of the analysis some attention is given to alternativeincome concepts which might be used in an incidence study.
In addition to the overall distributions of before- and after-taxincome, we also examine and compare the incidence of individualtaxes for specific subgroups of the population. And since weinterpret individual taxes broadly to include transfer paymentsreceived from the government as well as taxes paid, we assess theextent to which both individual taxes and transfers affect thedistribution of income in our analysis.
METHODOLOGY AND DATA
The results presented are all based on the Brookings MERGEFile which contains demographic and financial data for a sample•of 72,000 families and single individuals in calendar year 1966.This file was created by combining information from the 1967Survey of Economic Opportunity (SEO) and data from the 1966Tax File, which contains income and tax information from federalindividual income tax returns filed for 1966. The basic unit ofanalysis in the MERGE File is the Census family or unrelatedindividual.
2 While it is possible that individuals might change their allocation of timebetween work and leisure because of the individual income tax, there are feweconomists who believe that this would be a significant factor. Therefore, wefollow the traditional practice of allocating personal income taxes directlyamong individuals on the basis of their incomes under the assumption that weneed not take account of tax-induced changes in the distribution ofbefore-tax income. In the case of employment taxes, even though there is notunanimous agreement as to the incidence of such levies, we follow theprevalent modern practice of allocating both the employer and employeeshares of these taxes among persons on the basis of their compensation fromearnings. For a detailed discussion of the reasoning and various views on thissubject, see John A. Brittain, The Payroll Tax for Social Security (Wash-ington, D.C.: Brookings Institution, 1972).
A detailed description of the MERGE File and how it was created isgiven in Benjamin A. Okner, "Constructing a New Data Base from Existing
Individual Taxes and Distribution of Income 47
Since all the calculations are based on microunit data in theMERGE File, the methodology used here differs considerablyfrom previous empirical tax burden studies. In the past, individualtaxes were allocated to broad income classes on the basis of a largenumber of statistical series which were used as proxies for the taxdistributions.4 The major disadvantage of such methodology isthat it requires that taxes be distributed on the basis of the averageincome and behavior of all households in a particular income class,rather than on the basis of the income and behavior of theindividual microunits in the class.
Although we could not make all the distinctions that arerelevant for estimating tax liabilities, the MERGE File provides uswith a very rich source of information for this purpose. Amongthe characteristics that are particularly important for estimatingtax payments are sources of income, marital status and familycomposition, consumption patterns, and home ownership. Sincethis information is available for each unit in the file, whenever it isnecessary to make assumptions about the economic behavior ofhouseholds, we are not limited to a single assumption for allfamilies in a given income class. This frees us from the uniformityassumption which has been the hallmark of all past studies.
In addition to this major improvement in methodology, theMERGE File permits us to prepare tax burden distributions on thebasis of various alternative incidence assumptions and incomedefinitions. In the past, these were impossible because of the sheermagnitude of the computational job. As illustrated below, the newfile, along with present electronic computer capabilities, gives usgreat flexibility in this respect.
BEFORE-TAX DISTRIBUTION OF INCOME
There is no single concept of income that is acceptable anduseful for all analytical purposes. However, for analyzing theincidence of taxes it is clearly inappropriate to compare taxpayments with income subject to tax; we are interested in acomparison between taxes paid and total incomes. To provide thistype of information, we adopted a comprehensive income defini-
Microdata Sets: The 1966 MERGE File," Annals of Economic and SocialMeasurement 1 (July 1 972):325-42.
The classic study along these lines is by Richard A. Musgrave and others,"Distribution of Tax Payments by Income Groups: A case Study for 1948,"National Tax Journal 4 (March 195 1):11-54.
48 BenjaminA. Okner
tion that is intended to correspond as closely as is practical to aneconomic concept of income, viz., consumption plus tax paymentsplus (or minus) the net increase (or decrease) in the value of assetsduring the year. This concept, called family income (Fl), is thesum of national income (as defined in the National IncomeAccountS)5 plus transfer payments plus accrued gains on farmassets and nonfarm real estate6 In keeping with the nationalincome concept, Fl includes corporation income before tax. Thisprocedure has the advantage not only of consistency but also ofproviding a complete account of the accrued income claims of thehousehold sector. Retained earnings of corporations, which areincluded in family income, may be regarded as an approximationof accrued capital gains on corporate stock during the year.7 Flincludes only income which accrues directly to families andindividuals, and thus excludes the income received by fiduciariesand persons in the institutional population not represented in theSEO File.
Other income concepts that might be employed are moneyfactor income, Census money income, and total money receipts.The relationship among these four concepts is illustrated in Table1. Money factor income includes the $484 billion moneyincome received by individuals from production. To this we addwage supplements, net imputed rent, accrued capital gains on farmassets and nonfarm real estate, corporate retained earnings and the
The major departure from the official definition of income is theomission of interest imputed to individuals for the services rendered to themby the banking system.
6 For a detailed description, see Benjamin A. Okner, "Adjusted FamilyIncome: Concept and Derivation," Brookings Technical Working Paper II, forthe Distribution of Federal, State and Local Taxes Research Program, rev.,processed (Washington, D.C.: Brookings Institution, August 1972), which isavailable on request.
We used this approximation because the annual fluctuations in the valueof corporate stock are very large and even three- to five-year averages may notgive an adequate representation of accrued capital gains. Martin J. Bailey andMartin David have shown that over very long periods of time, capital gains oncorporate securities are roughly equal to retained earnings. See Martin J.Bailey, "Capital Gains and Income Taxation," in Arnold C. Harberger andMartin J. Bailey, eds., The Taxation of Income from Capital (Washington,D.C.: Brookings Institution, 1969), pp. 15-26; and Martin David,AlternativeApproaches to Capital Gains Taxation (Washington, D.C.: Brookings Institu-tion, 1968), pp. 242-46.
Individual Taxes and. Distribution of Income 49
TABLE I Comparison of Various Income Concepts, 1966
(billions of dollars)
Item Amount
Money factor incomea 484
Plus: Wage supplementsb 41
Net imputed rent 11
Retained corporate profits 22Corporation income tax 26Accrued gains on farm assets and nonfarm real estate 37Interest on life insurance policies 6
Subtotal 142
Equals family income before transfers 627
Less total nonmoney income —142
Plus: Private pensions 6
Civilian and military government pensions 6Other income 7
Subtotal —124
Equals Census money income before transfers 503Plus realized gains on asset sales 19
Equals total money receipts before transfers 523
Memorandum: transfer 34
a Money factor income is the sum of wages and salaries, interest, dividends, rents androyalties, and farm and nonfarm proprietors' income.
b Wage supplements mclude employer contributions to private pension and welfarefunds, Social Security, workmen's compensation, unemployment insurance, and civiliangovernment retirement systems.
C Transfer payments include Social Security benefits, public assistance, veterans'disability compensation and pensions; workmen's compensation, and unemploymentinsurance payments received by individuals.
corporation income tax allocated to individual stockholders,8 andinterest on life insurance policies to obtain total Fl beforetransfers of $624 billion. Since the Census income concept doesnot include nonmoney items, it is necessary to subtract the $141billion of such income from Fl before transfers and then addreceipts from private pensions and civilian government and
8 This excludes the portion of the tax attributable to fiduciaries andorganizations not represented in the SEO universe.
50 Benjamin A. Okner
military retirement pay to derive the $502 billion of Censusmoney income before transfers.9 Finally, because the Bureau ofthe Census does not include profits from the sale of assets in itsincome concept, we add $19 billion of realized capital gains toderive $521 billion of total money receipts before transfers.
These four different concepts vary considerably in theircoverage. Family income is the most comprehensive concept,while money factor income is least inclusive. And as might beexpected, the before tax and transfer distribution of income isquite different, depending on which concept is used. InTable 2 we present the distribution of income by size class undereach of the four different definitions. (In Table 2, the incomeclasses are defined in terms of each of the specific incomedefinitions; therefore, the same family units are not included inthe same income class under the various definitions.)
In Table 3, we show the same distributions by populationquintiles under the different income definitions. Based on the datain Table 3, it is obvious that one can reach very differentconclusions about the distribution of income depending uponwhat income concept is used. For example, under the Flconcept, the lowest quintile includes families with incomes under$2,799; under the least comprehensive money factor income(MFI) concept, the lowest quintile includes those with incomes ofless than $2,368, At the other end of the income distribution, thetop quintile under the Fl concept is comprised of families withincomes of $14,564 and over, whereas under the MFI definition,the highest quintile starts at $12,446. The absolute differences inincome increase even more at the very highest levels; the top onepercent of units are those with $44,318 of MFI while the toppercentile under the Fl concept begins at $50,000.
As measured by the Gini coefficient,1° the distributions also
Under our definition, only social insurance benefits financed by payrolltaxes are treated as transfer payments. Benefits paid under governmentretirement programs are treated like private pensions and annuities and,therefore, are excluded from transfers. Government contributions, if any, tothese programs are considered income during the current year and benefitpayments received are viewed as representing only a change in the form ofasset holding by individuals (i.e., cash is increased and a prepaid insuranceasset is reduced).
1 ° The Gini coefficient of inequality is a statistical measure of overallequality or inequality in the distribution of income. Pictorially, it is equal tothe ratio of the area between the Lorenz curve and the line of equaldistribution to the entire area below the line of equal distribution. The value
Individual Taxes and Distribution of Income 51
TABLE 2 Distribution of Income Before Taxes and Transfers UnderVarious Income Definitions, by Income Class, 1966
(millions of dollars and percent)
Income Money Census TotalClassesa Factor Family Money Money($000) Income Income Income Receipts
Under 3b 10,724 10,784 11,386 11,2863- 5 28,053 23,563 29,578 29,4505- 10 151,516 133,594 152,566 151,284
10- 15 138,406 162,444 139,953 139,28615- 20 61,959 101,637 68,538 69,924
20- 25 24,304 50,054 26,679 28,017
25- 50 43,712 72,894 46,784 51,78250- 100 15,937 28,881 16,665 20,112
100- 500 8,153 29,470 8,643 14,224
500-1,000 655 4,854 681 2,0511,000 and over 632 6,658 645 3,178
All classes 484,050 624,833 502,118 520,594
Percentage Distributions
Under3b 2.2 1.7 2.3 2.23- 5 5.8 3.8 5.9 5.75- 10 31.3 21.4 30.4 29.1
10- 15 28.6 26.0 27:9 26:815- 20 12.8 16.3 13.6 13.4
20- 25 5.0 8.0 5.3 5.425- 50 9.0 11.7 9.3 9.950- 100 3.3 4.6 3.3 3.9
100- 500 1.7 4.7 1.7 2.7500-1,000 0.1 0.8 0.1 0.4
1,000 and over 0.1 1.1 0.1 0.6
All classes 100.0 100.0 100.0 100.0
NOTE: Details may not add to totals because of rounding.a The income class is defined in terms of the income distribution for each of the
concepts.Includes negative incomes.
of the Gini coefficient varies between 0 (indicating perfect equality) and I(indicating perfect inequality). A decrease in the value of the coefficientsignifies a more equal distribution; an increase signifies a more unequaldistribution.
Mon
ey F
acto
r In
com
eFa
mily
Inc
ome
Cen
sus
Mon
eyIn
com
eT
otal
Mon
ey R
ecei
pts
Perc
ent
Perc
ent
Perc
ent
Perc
ent
ofof
ofof
Popu
latio
nIn
com
eIn
com
eIn
com
eIn
com
eIn
com
eIn
com
eIn
com
eIn
com
eQ
uint
ileR
ange
Rec
eive
dR
ange
Rec
eive
dR
ange
Rec
eive
dR
ange
Rec
eive
d
Low
est 5
thU
nder
$2,
368
1.41
Und
er $
2,79
91.
90U
nder
$2,
482
2.05
Und
er $
2,50
22.
01Se
cond
5th
2,36
8- 5
,389
9.50
2,79
9- 6
,534
9.30
2,48
2- 5
,582
9.79
2,50
2-5,
630
9.54
Mid
dle
5th
5,38
9- 8
,383
17.0
76,
534-
9,9
8216
.07
5,58
2- 8
,567
17.0
05,
630-
8,63
416
.53
Four
th 5
th8,
3834
2,44
624
.57
9,98
2-14
,564
23.2
98,
567-
12,7
3624
.22
8,63
4-12
,884
23.7
2H
ighe
st 5
th12
,446
and
ove
r47
.39
14,5
64 a
nd o
ver
49.4
412
,736
and
ove
r46
.94
12,8
84 a
nd o
ver
48.3
0T
op 5
per
cent
19,5
76 a
nd o
ver
20.2
324
,459
and
ove
r23
.48
19,8
48 a
nd o
ver
20.0
520
,714
and
ove
r21
.80
Top
1 p
erce
nt44
,318
and
ove
r7.
5650
,000
and
ove
r11
.02
44,5
65 a
nd o
ver
7.42
46,1
52 a
nd o
ver
9.01
Gin
i coe
ffic
ient
of
ineq
ualit
ya.4
601
.474
6.4
489
.462
2
a T
he G
ini c
oeff
icie
nt o
f in
equa
lity
is a
sta
tistic
al m
easu
re o
f ov
eral
l equ
ality
in th
e di
stri
butio
n of
inco
me.
It m
ay v
ary
betw
een
0(i
ndic
atin
g pe
rfec
t equ
ality
) an
d 1
(ind
icat
ing
perf
ect i
nequ
ality
). A
dec
reas
e in
the
valu
e th
eref
ore
sign
ifie
s a
mor
e eq
ual a
lter-
tax
dist
ribu
tion
of in
com
ean
d a
mor
e pr
ogre
ssiv
e ta
x st
ruct
ure.
TA
BL
E 3
Shar
es o
f In
com
e B
efor
e T
axes
Def
initi
ons,
196
6an
d T
rans
fers
Rec
eive
d by
Eac
h Fi
fth
of th
e Po
pula
tion
Und
er V
ario
us I
ncom
e
Individual Taxes and Distribution of Income 53
differ substantially in the degree of inequality exhibited under thevarious income definitions. Census money income is distributedamong families most nearly equal (Cmi coefficient = .4489). Inorder of decreasing equality (increasing inequality), this isfollowed by money factor income (Gini coefficient = .460 1); totalmoney receipts (Gini coefficient = .4622); and finally, familyincome (Gini coefficient = .4746). All of the preceding coeffi-cients refer to the distributions of income exclusive of transferpayments.
DISTRIBUTION OF INDIVIDUAL TAXES AND TRANSFERS
As has been indicated, the analysis is confined to the effects ofpersonal income and employment taxes on the distribution ofindividual income. Since in this paper, we are dealing only withindividual taxes, the amount of the corporation income taxallocated to each family is excluded from family income in all theeffective tax-rate tables. However, in each table, families areclassified by the amount of total income before taxes andtransfers, i.e., total Fl less transfer payments.
The income taxes include federal and state and local taxes.Payroll taxes include employee and employer Social Securitycontributions plus employer contributions for unemploymentinsurance and workmen's compensation. Other employer socialinsurance contributions—such as those for pension and healthfunds—are not considered taxes and are excluded from theanalysis. Transfer payments include all benefits paid to individualsunder government programs, regardless of whether they arefinanced by general revenue or by payroll taxes.'1
Personal Income Taxes
The MERGE File contains the federal individual income taxreported by each taxpaying unit in 1966, so there is no need toallocate these taxes among families in the file.
State and local income taxes reported as itemized deductions ontax returns amounted to approximately 70 percent of the totalcollections reported for 1966. However, for units that did not
1 1 Business transfer payments, as defined in the National IncomeAccounts, are not included in our transfer payment (or income) definition.
54 Benjamin A. Okner
itemize their deductions, we estimated the state and local incometax liability on the basis of income. family size, and place ofresidence. 12
Total personal income tax collections (less refunds) in theNational Income Accounts were $64.0 billion in 1966; this wascomprised of $58.6 billion in federal collections and $5.4 billionin state and local government collections. The $55.4 billion offederal tax collections in the MERGE File amounts to about 90percent of the national income amount. We did not attempt toadjust the MERGE figures to the national income total andaccepted the amount reported.'3 State and local income taxesallocated to MERGE File units were $5.4 billion. The totalamount of personal income taxes allocated among families in thefile is therefore $60.8 billion.
Total federal income taxes amount to 9.2 percent of income,while total state and local income taxes equal 0.9 percent of Fl.These taxes as a percentage of total Fl in each income class areshown in Table 4•14 As can be seen, both federal and state andlocal income taxes are progressive throughout most of the incomedistribution. However, for most families, total personal income taxrates are quite low. The federal income tax never exceeds 20.7percent of total income, and the highest effective rate of state andlocal income taxes is only 1 .9 percent. The effective tax rate is 10percent or less for all families with incomes below $20,000 andexceeds 20 percent only for those with incomes of $100,000 andabove. The highest effective rate of tax—22.7 percent—is reachedin the $500,000 to $1 million Fl class; beyond this income level,the effective income tax rate declines, because nontaxable incomeis highly concentrated among those with incomes at the very topof the income distribution.
1 2 The last criterion is needed because not all states have an individualincome tax. While it is used extensively, in 1966 there were still seventeenstates that did not levy a personal income tax.
1 3 An unknown, but probably small, part of the difference can beattributed to the fact that the SEO population differs from the nationalincome covereage.
1 Similar tables, with effective income tax rates based on the otherincome concepts, are presented in Appendix Tables A.! to A.3. In order toaid the reader, families are classified by family income less transfers in all, theAppendix Tables so that the effects of the income definition changes for thesame families may be followed.
Individual Taxes and Distribution of Income 55
TABLE '4 Effective Rates of Federal and State and Local IndividualIncome Taxes,a by Family Income Classes, 1966
(percent)
Family Federal State and Local TotalIncome Before Individual Individual Individual
Transfers In come Income Income($000) Tax Taxes Taxes
0- 3b 2.7 0.2 3.03- 5 4.6 0.4 5.05- 10 6.7 0.6 7.2
10- 15 8.1 0.8 8.915- 20 9.1 0.9 10.020- 25 9.9 1.1 11.025- 50 11.4 1.2 12.650- 100 17.3 1.7 19.0
100- 500 19.6 1.9 21.5500-1,000 20.7 1.9 22.7
1,000 and over 19.0 1.8 20.8
All classes 9.2 0.9 10.2
NOTE: Details may not add to totals because of rounding.a Effective tax rates are calculated on the basis of family income before transfers,
excluding the amount of corporation income tax allocated to families in the MERGEFile.
b Excludes families with negative incomes.
It is well known that average effective rates, such as thoseshown in Table 4, often obscure large variations in taxes paid bydifferent kinds of families within the same income class. Thedifferences between those who derive their incomes primarilyfrom wages and those who are primarily recipients of propertyincome were reported in a recent paper by Joseph A. Pechman.'5There are a very large number of other population subgroups thatmight be examined: two of these which are of particular interestare families of different size and those headed by aged andnonaged persons. The effective income tax rates for thesesubgroups are shown in Table 5.
1 See Joseph A. Pechman, "Distribution of Federal and State IncomeTaxes by Income Classes," Papers and Proceedings of the Thirtieth AnnualMeeting of the American Finance Association, Journal of Finance 27, no. 2(May 1972):179-91. Brookings Reprint 234.
TA
BL
E 5
Effe
ctiy
eR
ates
of
Indi
vidu
al in
com
e T
axes
a by
Age
of
Fam
ily H
ead
and
Size
of
Fam
ily, b
y Fa
mily
inco
me
Cla
sses
, 196
6(p
erce
nt)
Fam
ilyIn
com
eB
efor
eT
rans
fers
($00
0)
All
Fam
ilies
Non
aged
Fam
ilies
bA
ged
Fam
ilies
b
All
Size
s1
2
Fam
ily S
ize
All
5#Si
zes
Fam
ily S
ize
All
5+Si
zes
Fam
ily S
ize
34
12
34
12
34
5+
3C3.
04.
12.
52.
0L
8L
i3.
95.
53.
72.
31.
91.
91.
72.
01.
70.
71.
1d
3-5
5.0
8.5
4.3
4.2
2.9
L6
5.6
10.0
5.4
4.4
2.7
1.5
3.1
3.8
2.7
2.8
4.8
2.2
5-10
7.2
11.5
8.2
7.4
6.1
3.9
7.4
12.4
8.9
7.5
6.1
3.9
5.5
4.9
5.7
6.0
5.4
4.5
10-
158.
913
.810
.69.
78.
26.
79.
015
.110
.99.
78.
26.
78.
27.
48.
38.
88.
76.
115
-20
10.0
12.6
11.4
10.4
9.8
8.6
10.1
14.2
12.0
10.7
9.8
8.6
8.3
8.2
7.8
8.6
9.8
8.5
20-
2511
.012
.712
.711
.110
.89.
911
.316
.813
.511
.610
.99.
97.
65.
07.
76.
99.
210
.525
-50
12.6
11.4
12.5
13.4
12.8
12.5
13.1
16.1
13.8
13.7
12.8
12.5
8.0
6.1
7.2
10.3
12.6
6.0
50-
100
19.0
21.7
17.1
18.9
20.2
20.0
20.2
23.3
20.7
19.0
20.3
20.0
4.6
9.6
4.3
d—
100-
500
21.5
16.4
22.0
24.7
24.1
20.3
21.6
16.4
22.1
24.7
24.1
20.3
10.3
—10
.4—
——
500-
1,00
022
.719
.423
.123
.725
.923
.022
.723
.123
.725
.923
.0—
——
——
—
1,00
0an
d ov
er20
.819
.820
.323
.821
.021
.720
.819
.820
.323
.821
.021
.7—
——
——
All
clas
ses
10.2
11.6
11.3
10.5
9.8
8.7
10.5
13.3
12.5
10.7
9.8
8.7
6.2
5.0
5.8
7.7
8.5
6.6
a E
ffec
tive
tax
rate
s ar
e ca
lcul
ated
on
the
basi
s of
fam
ily in
com
e be
fore
tran
sfer
s, e
xclu
ding
the
amou
nt o
f co
rpor
atio
n in
com
e ta
x al
loca
ted
tofa
mili
es in
the
ME
RG
E F
ile.
b Fa
mili
es h
eade
d by
an in
divi
dual
age
64
or u
nder
are
con
side
red
nona
ged;
thos
e he
aded
by
an in
divi
dual
age
65
or o
ver
are
clas
sifi
ed a
s ag
ed.
C E
xclu
des
fam
ilies
with
neg
ativ
e in
com
es.
Les
s th
an h
alf
of 1
per
cent
.
Individual Taxes and Distribution of Income 57
For aged families, income taxes as a percentage of income aresubstantially below the rates paid by the nonaged at all incomelevels. On the average, the aged pay income taxes at about 60percent of the rates paid by families headed by an individual underage 65.
While the same general pattern of lower tax rates is found byincome class for each family-size group, we see a very differentoverall pattern of effective income tax rates among the aged andnonaged as family size increases. For nonaged families, effectivetax rates fall as family size increases, whereas just the oppositeoccurs among aged families. This occurs because a large familyheaded by a person age 65 or over is very likely comprised of thehead plus other, younger, family members still in the labor force,whereas an aged one-person family is likely to be a widow orwidower. In general, as family size increases among aged families,there are likely to be more earners, larger incomes, and thereforehigher tax payments.'6 On the other hand, among nonagedfamilies, larger family size more typically represents more depen-dents and a lower likelihood of additional earners other than thefamily head (or head and spouse) than is the case among theaged.'7
Employment Taxes
Employer and employee payroll taxes amounted to $31.8billion in 1966. As indicated above, the employer payroll taxes aredefined to include only the portion of social insurance contribu-tions for Social Security and unemployment insurance in theNational Income Accounts. In addition, employer workmen'scompensation costs, which are excluded from wage supplements inthe National Income Accounts, are included here as an employerpayroll tax.
Since neither the employee nor employer payroll tax data wereavailable from the SEO or Tax Files used in constructing theMERGE File, these amounts were allocated to workers on the
1 6 The average income for aged single persons is about $2,400. This risessharply as family size increases up to an average of almost $8,900 for agedfamilies with five or more persons.
The average income for nonaged single persons is about $5,800. Thisrises to about $11,200 for two-person families and then remains in the$11,500 to $12,600 range for all other family sizes.
58 Benjamin A. Okner
basis of their earnings, industry and occupation, and the statutoryrequirements in effect in 1966.18 Employer Social Securitycontributions for self-employed individuals were available directlyfrom the federal individual income-tax data.
Since we accept the assumption that employer payroll taxes areultimately borne by employees, we follow the national incomeprocedure and include such levies as part of employee compensa-tion in Fl. The effective rate of such taxes is then correctlycomputed as the ratio of the tax to income before tax.'9 There islittle disagreement over who pays the employee share of thepayroll tax, and using the standard assumption, it was allocatedamong wage and salary earners. The effective payroll tax ratesunder these assumptions are shown in Table 6. As in the case oftotal individual income taxes, these are shown for all families andalso by family size for aged and nonaged families.
The overall pattern of employment taxes shows the expectedregressive pattern by income class. Although for both aged andnonaged families, tax rates ultimately fall as income rises, theeffective tax rate is lower for those with incomes under $3,000than it is at incomes of $3,000 to $5,000. Since the payroll taxesare essentially proportional to earnings in covered employment(up to the $6,600 maximum of taxable wages in 1966), thepattern of effective rates indicates that families in the lowestincome class have a low proportion of covered earnings to totalincome. For the aged, this occurs because a large proportion ofincome is derived from property and is not subject to the payrolltax. Among nonaged families, low effective payroll tax rates at thebottom of the income scale result primarily from working in
1 8 For details on the allocation process, see Benjamin A. Okner, "TheImputation of Missing Income Information," Technical Working Paper III, forthe Distribution of Federal, State and Local Taxes Research Program,processed (Washington, D.C.: Brookings Institution, April 1971), which isavailable on request.
1 Under this incidence assumption, employer payroll taxes should beadded to reported money earnings in order to correctly assess the impact ofthese taxes under other income definitions. Alternative incidence assumptionsare not included in this paper, since the distributional effects tend to becomevery complex. For example, if it is assumed that employer taxes are shiftedforward in the form of higher commodity prices, it is necessary to move fromthe sources to the uses of income side in the analysis. This, in turn, requiresthat net national product (which includes indirect business taxes) be used asthe tax base instead of the "national income" concept used here.
TA
BL
E 6
Eff
ectiv
e R
ates
of
Em
ploy
men
t Tax
esa
by A
ge o
f Fa
mily
Hea
d an
d Si
ze o
f Fa
mily
, by
Fam
ily I
ncom
e C
lass
es, 1
966
(per
cent
)
Fam
ilyIn
com
eB
efor
eT
rans
fers
($00
0)
All
Fam
ilies
Non
aged
Fam
ilies
!'A
ged
Fam
ilies
!'
All
Size
s
Fam
ily S
ize
All
5+Si
zes
Fam
ily S
ize
All
Size
s
Fam
ily S
ize
12
34
12
34
5+1
23
45+
0-3C
5.4
5.1
4.4
6.2
7.7
8.8
7.6
7.5
7.2
7.0
8.0
9.0
2.3
1.5
2.6
3.5
6.5
5.8
3-5
7.1
6.9
6.3
7.5
8.1
8.0
8.2
8.4
8.3
7.9
8.2
8.0
3.3
2.2
3.2
5.3
6.7
7.1
5-10
7.3
6.8
6.9
7.6
7.5
7.6
7.5
7.3
7.6
7.7
7.5
7.6
4.8
3.1
4.6
6.8
7.3
7.4
10-
156.
24.
46.
46.
86.
26.
26.
45.
06.
66.
86.
26.
24.
61.
34.
46.
06.
97.
2
15-
205.
23.
05.
25.
35.
55.
35.
43.
45.
55.
55.
45.
33.
91.
93.
04.
18.
06.
820
-25
4.6
1.9
3.8
5.3
4.9
4.7
4.7
2.9
4.0
5.4
4.8
4.7
3.0
d2.
43.
75.
94.
425
-50
2.8
1.0
1.8
3.3
3.4
3.1
2.9
1.4
2.0
3.2
3.3
3.1
2.0
0.5
1.1
3.6
7.0
4.1
50-
100
0.9
0.5
0.7
0.8
1.2
1.0
0.9
0.6
0.8
0.8
1.2
1.0
0.3
d0.
4d
2.2
100-
500
0.3
d0.
30.
40.
50.
40.
30.
10.
30.
40.
50.
4d
0.4
—
500-
1,00
0d
d0.
10.
10.
10.
1d
d0.
10.
10.
10.
1—
1,00
0and
over
dd
dU
dU
dU
dd
dd
All
clas
ses
5.3
4.9
4.9
5.8
5.5
5.3
5.5
5.7
5.2
5.8
5.5
5.3
3.6
1.8
3.2
5.0
7.0
6.4
a E
ffec
tive
tax
rate
s ar
e ca
lcul
ated
on
the
basi
s of
fam
ily in
com
e be
fore
tran
sfer
s, e
xclu
ding
the
amou
nt o
f co
rpor
atio
n in
com
e ta
x al
loca
ted
tofa
mili
es in
the
ME
RG
E F
ile.
b Fa
mili
es h
eade
d by
an
indi
vidu
al a
ge 6
4 or
und
er a
re c
onsi
dere
d no
nage
d; th
ose
head
ed b
y an
indi
vidu
al a
ge 6
5 or
ove
r ar
e cl
assi
fied
as
aged
.C
Exc
lude
sfa
mili
es w
ith n
egat
ives
inco
mes
.d
Les
s th
an h
afi o
f 1
perc
ent.
60 Benjamin A. Okner
occupations not covered by Social Security and the otherprograms, which tends to lower the proportion of taxable wages tototal income for these units. For nonaged families with incomes of$3,000 to $5,000, the effective tax rate on family incomeis about 8 percent, and this falls steadily as income rises above thatlevel. There is very little difference in this pattern among nonagedfamilies of different sizes. The burden of employment taxesamong aged families is generally low, because a much smallerproportion of such units are in the labor force and subject to thepayroll levies.
Because larger families in this category tend to have moreearners, we find that effective payroll tax rates rise as family sizeincreases for families headed by an aged person. Among nonagedfamilies in the $10,000 to $20,000 income range, there is a sharprise in effective payroll taxes paid by two-person families ascompared with single individuals. This undoubtedly represents amove from single-earner to two-earner status for units in thisincome range.
Total Individual Taxes
When we examine the combined effect of the regressiveemployment taxes and the progressive income taxes, we find thatthe overall pattern of total tax burdens is slightly progressive. Theeffective tax rates by income classes for both taxes combined aregiven in Table 7. Those figures suggest that the progressivityinvolved in the combined data comes almost totally from theeffect of the individual income tax near the top of the incomedistribution.
For the vast bulk of families, the combined effect of theindividual income and employment taxes is pretty much propor-tional with respect to income. For example, among familiesheaded by a nonaged individual, the combined effective tax rate isbetween 1 5 percent and 1 6 percent for all income levels between$5,000 and $50,000. This group comprises about 68 percent of allfamily units and receives almost 80 percent of all income (beforetaxes and transfers).
Families headed by a person age 65 and over pay lower tax rateson the average, but again there is not a great deal of variation inthe combined effective rates of tax. Among the aged families, theeffective tax rate ranges from about 10 percent to 13. percent of
Individual Taxes and Distribution of Income 61
TABLE 7 Combined Effective Rates of Individual Income and Employ-inent Taxesa by Age of Family Head and Family IncomeClasses, 1966
(percent)
FamilyIncome Before
Transfers($000)
AllFamilies
Nonaged Aged
0- 3C 8.4 11.5 4.03- 5 12.1 13.8 6.45- 10 14.5 15.0 10.3
10- 15 15.2 15.3 12.815- 20 15.2 15.5 12.220- 25 15.6 16.0 10.625- 50 15.5 16.0 9.950- 100 19.8 21.1 4.9
100- 500 21.8 21.9 11.2500-1,000 22.7 22.7 —
1,000 and over 20.8 20.8 —
All classes 15.4 16.0 9.8
a Effective tax rates are calculated on the basis of family income before transfers,excluding the amount of corporation income tax allocated to families in the MERGEFile.
b Families headed by an individual age 64 or under are considered nonaged; thoseheaded by an individual age 65 or over are classified as aged.
C Excludes families with negative incomes.
income for all units between the $5,000 and $50,000 incomelevels. Since there are no aged families at the very top of theincome distribution in the MERGE File sample, the small degreeof progressivity for the aged all comes from the lower taxes paidby those near the bottom of the income distribution.
Transfer Payments
Since transfer payments have a direct impact on the distributionof income available for private use, it seems clear that they shouldalso be included in this analysis. Our definition of transfers is quitesimilar to that used in the National Income Accounts. The majordifferences are that we do not count either civilian or government
62 Benjamin A. Okner
retirement receipts as transfers but we do include workmen'scompensation benefits in transfer income.
Under this definition, total transfers to individuals amounted to$34 billion in 1966. This was primarily comprised of the $21.5billion of Social Security benefits paid; it also included unemploy-ment insurance, public assistance, veterans' disability compensa-tion and pensions, and workmen's compensation receipts.
Since the most relevant distinctions among families withtransfer receipts are age and income, we omit the family-sizeclassification in showing the effect of transfer payments in Table8. Transfer payments have their greatest impact on low-incomefamilies and especially the aged. In fact, for aged persons withincomes below $3,000, transfers average more than double theamount of income from production. Transfer income amounts to
TABLE 8 Transfer Payments as aPercentage of Incomea by Age of FamilyHead and Family Income, 1966
(percent)
FamilyIncome Before
Transfers($000)
AllFamilies
NonagedFam
0- 3C 124.2 74.4 203.03- 5 17.9 11.5 39.65- 10 5.1 3.5 19.9
10- 15 2.2 L6 10.915- 20 1.8 1.3 9.020- 25 1.4 1.1 5.225- 50 0.9 0.6 4.250- 100 0.2 d 2.6
'100- 500 •d 1.9
500-1,000 —
1,000 and over —
All classes 5.7 3.0 333
a Percentages are calculated on the basis of family income before transfers, excludingthe amount of corporation income tax allocated to families in the MERGE File.
b Families headed •by an individual age 64 or under are considered nonaged; those'headed by an individual 65 or over are classified as aged.
C Excludes families with negative incomes.d Less than half of 1 percent.
Individual Taxes and Distribution of Income 63
almost 40 percent of before-transfer income for the aged withincomes of $3,000 to $5,000 and close to 20 percent of Fl forthose with incomes of $5,000 to $10,000. As expected, transfersare much less important in influencing the distribution of incomeamong nonaged families except. at the very bottom of the incomescale. Transfer payments are a very small proportion of totalincome for nonaged families with incomes of $5,000 or more.
DISTRIBUTION OF INCOME AFTERTAXES AND TRANSFERS
In this section, we combine the partial results discussed aboveto assess the overall impact of individual taxes and transfers on thedistribution of before-tax income. The most expeditious way tosummarize the large amount of information already presented is interms of Lorenz curves showing the various income distributions.The before-tax and after-tax and transfer Lorenz curves for allfamilies are shown in Chart 120 and the Gini coefficients andpercentage reductions in the areas of inequality for each of thepopulation subgroups examined are given in Table 9•21 (Thecumulative distributions of before- and after-tax income for allfamilies and for each of the population subgroups are given inAppendix Tables A.5 to A.7.)
For all families, the Gini coefficient computed on the basis ofthe before-tax distribution of income is .459 5; the coefficient forincome after transfer payments is .4 155; for income less transfersand income taxes, it is .3959; for income less transfers andemployment taxes combined, it is .4200; and the Gini coefficientfor the income distribution after both taxes and transfers is .3998.Translating these figures into more commonly used terms, they
20 Before-tax income for the distribution in Chart 1 is equal to F! beforetransfers and excluding the amount of corporation income tax allocated toeach family in the MERGE File. This is also the basis used for computing thebefore-tax and transfer Gmi coefficients in Table 9.
2 1 The percentage reduction in the area of inequality is equal to the ratioof the area between the before- and after-tax Lorenz curves (B) to the totalarea of inequality, i.e., the area between the line of equal distribution and thebefore-tax Lorenz curve (A). This ratio can be computed directly from theGini coefficients associated with the before- and after-tax Lorenz curves. Ifthe before-tax Gini coefficient is equal to G, and the after-tax Gini coefficientequals G', the percentage reduction in the area of inequality equal to(G - G')/G.
64 Benjamin A. Okner
Cumulative100
90
80
70
60
50
40
30
20
1 0
00 10 20
Cumulative percentage of families90 100
CHART I: Comparison of the Distribution of Family Income Before andAfter Individual Taxes and Transfers, All Families, 1966
indicate that in the aggregate, income taxes are progressive;employment taxes are regressive; the totalemployment taxes is progressive; and that transfer payments arevery progressive.
In general, income before taxes and transfers is more equallydistributed among nonaged families than among those headed bysomeone age 65 or over. The group with the most unequaldistribution of before-tax income consists of aged single mdi-viduals; they are closely followed by aged couples. At the otherend of "the equality scale" are the "standard" four-person familiesheaded by a person under age 65.
Based on the changes in the area of inequality shown in Table 9,
percentage of income
30 40 50 60 70 80
of income and
Individual Taxes and Distribution of Income 65
TABLE 9 Gini Coefficients for the Distributions of Income Before Taxesand Transfers, Income After Transfers, and Income After Taxesand Transfers, 1966
Gini Percentage Change in Areaof Due to..
Income IncomeIncomeAfter Individual
Population Before After Taxes and Taxes andGroup Taxesc Transfers Transfers Transfers Transfers
All families .4595 .4155 .3998 9.6 13.0Nonaged famiiesd .4099 .3886 .3774 5.2 7.9
I person .5102 .4717 .4570 7.5 10.42 persons .43 10 .4059 .3919 5.8 9.13 persons .3628 .3427 .3297 5.5 9.14 persons .3385 .3255 .3126 3.8 7.65+ persons .3639 .3433 .3277 5.2 9.9
Aged famiiesd .6278 .4573 .4367 27.2 30.41 person .6799 .4263 .4097 37.3 39.72 persons .5842 .4129 .3944 29.3 32.53 persons .4598 .3744 .3595 18.6 21.84 persons .4869 .3916 .3647 19.6 25.15+ persons .4231 .3470 .3309 18.0 21.8
a The Gini coefficient of inequality is a statistical measure of overall equality orinequality in the distribution of income. It may vary between 0 (indicating perfectequality) and 1 (indicating perfect inequality). A decrease in the value therefore signifiesa more equal after-tax distribution of income and a more progressive tax structure.
b The percentage reduction in the area of inequality is equal to the ratio of the areabetween the before-tax and the after-tax (after-transfer) Lorenz curves to the total areaof inequality, i.e., the area between the line of equal distribution and the before-taxLorenz curve.
C Income before taxes is equal to family income less transfers, excluding the amountof corporation income tax allocated to families in the MERGE File.
d Families headed by an individual age 64 or under are considered nonaged; thoseheaded by an individual age 65 or over are classified as aged.
we find that transfer payments have a much greater effect on theafter-tax and transfer distribution of income than do tax pay-rnents. For all families, transfer payments account for aboutthree-quarters of the reduction in the area of inequality, whereastaxes account for one-fourth of the total change. Approximatelythe same proportions of total change are also attributable to the
66 Benjamin A. Okner
effects of transfers and taxes for one-person nonaged families. Fortwo- and three-person nonaged families, about 60 percent of thetotal reduction in inequality can be attributed to transfers and 40percent to taxes; the proportions are about 50 percent each fortaxes and transfers among larger nonaged families.
Transfer payments are extremely important in reducinginequality in the distribution of income among the aged. For suchunits, transfers account for a minimum of about 80 percent of thetotal reudction in the area of inequality (among four-personfamilies) and they are responsible for 94 percent of the totalchange among single individuals over age 65. It should also benoted that the total percentage changes in inequality between thebefore-tax and the after-tax and transfer distributions for agedfamilies are all substantially larger than they are for the nonagedgroup.
CONCLUSIONS
On the basis of the data presented here, it is clear that (1) thenet effect of direct federal taxes and transfers has an importantimpact on the distribution of individual incomes in the economy;and (2) of the two parts, transfers play a far more important rolein redistributing income among families than do taxes.2 2 Sincethere have been two federal income tax reductions since 1966, it ispossible that we are understating the redistributive effects of taxesin this analysis. However, there have also been significant increasesin public assistance and Social Security benefits (plus payroll taxincreases) during the period which would tend to offset some ofthe tax reduction effects. The data needed to assess the impact ofthese changes are not available, but I do not believe that the majorconclusions would be very different if these new features weretaken into account. On balance, I would guess that taxes nowaccount for a little bit more of the total redistribution, whiletransfers account for a slightly smaller degree of redistribution.Thus, if further income redistribution is an important national
22 The major finding of a study of overall tax burdens in 1966 was thattotal taxes—federal, state, and local—are proportional to income for almost90 percent of all U.S. families. (See Joseph A. Pechman and Benjamin A.Okner, Who Bears the Tax Burden [Washington, D.C.: Brookings Institution,1974]). As a result, the total tax system had a very small effect on the overalldistribution of income.
Individual Taxes and Distribution of Income 67
objective, we must either adopt changes that will increase theprogressivity of existing taxes.2 3 and/or expand transfer paymentsto individuals using financing arrangements which are not regres-sive.
APPENDIX
TABLE A.1 Effective Rates of Total Individual Income Taxes Based onCensus Money Incomea by Age of Head and Family IncomeClasses, 1966
(percent)
Family In comeBefore Transfers
($000)A!!
FamiliesNonagedFamilies1'
AgedFamiliesb
3c 3.0 4.2 1.63- 5 6,0 6.6 185- 10 8.6 8.8 6.8
10- 15 10.6 10.6 10.015- 20 11.8 12.0 10.220- 25 13.1 13.4 9.525- 50 15.6 16.0 11.550- 100 24.5 25.8 6.8
100- 500 33.8 33.9 18.6500-1,000 45.5 45.5 —
• 1,000 and over 47.4 47.4 —
All classes 12.1 12.6 7.4
a Effective tax rates computed on the basis of Census money income excludingtransfers.
b Families headed by an individual age 64 or under are considered nonaged; thoseheaded by an individual age 65 or over are classified as aged.
C Excludes families with negative income.
23 For ways to increase federal income tax progressivity, see Joseph A.Pechman and Benjamin A. Okner, "Individual Income Tax Erosion byIncome Classes," in The Economics of Federal Subsidy Programs, ACompendium of Papers submitted to the Joint Economic Committee, Part 1,General Study Papers, 92 Cong., 2 sess., 1972, pp. 13-40. Brookings Reprint230. For a discussion of measures that would reduce the regressivity ofpayroll taxes, see Joseph A. Pechman, Henry J. Aaron, and Michael K.Taussig, Social Security: Perspectives for Reform (Washington, D.C.:Brookings Institution, 1968), pp. 214-27.
68 Benjamin A. Okner
TABLE A.2 Effective Rates of Total Individual Income Taxes Based onMoney Factor Income, by Age of Head and Family IncomeClasses, 1966
(percent)
Family IncomeBefore Transfers
($000)All
FamiliesNonagedFamilies0
AgedFamilies'2
0- 3b 47 53 343- 5 6.5 6.9 4.85- 10 8.8 9.0 7.6
10- 15 10.8 10.8 10.715- 20 12.1 12.2 10.8 '
20- 25 13.4 13.7 10.125- 50 16.0 16.3 12.050- 100 24.9 • 26.3 7.0
100- 500 34.6 34.7 18.6500-1,000 46.4 46.4 —
1,000 and over 47.9 47.9 —
All classes 12.6 12.9 8.7
a Families headed by an individual age 64 or under are considered nonaged; thoseheaded by an individual age 65 or over are classified as aged.
b Excludes families with negative income.
Individual Taxes and Distribution of Income 69
TABLE A.3 Effective Rates of Total Individual Income Taxes Based onTotal Money by Age of Head and Family IncomeClasses, 1966
(percent)
Family IncomeBefore Transfers
($000)All
FamiliesNonagedFamiliesb
AgedFamiliesb
0 3C 2.3 40 1.5
3- 5 5.9 6.5 3.65- 10 8.5 8.7 6.3
10- 15 10.5 10.5 9.315- 20 11.6 11.8 9.620- 25 12.6 12.9 9.125- 50 14.9 15.2 10.850- 100 22.1 23.2 6.6
100- 500 27.3 27.4 18.6500-1,000 32.8 32.8 —
1,000 and over 32.2 32.2 —
All classes 11.7 12.1 7.0
a Effective tax rates computed on the basis of total money receipts excludingtransfers.
b Families headed by an individual age 64 or under are considered nonaged; thoseheaded by an individual age 65 or over are classified as aged.
C Exc'udes families with negative income.
TA
BL
E A
.4C
ombi
ned
Eff
ectiv
e R
ates
of
Indi
vidu
al I
ncom
e an
d E
mpl
oym
ent T
axes
a by
Age
of
Fam
ily H
ead
and
Size
of
Fam
ily,
by F
amily
Inc
ome
Cla
sses
, 196
6
(per
cent
)
Fam
ilyIn
com
eT
rans
fers
($00
0)
All
Fam
ilies
No
piag
ed F
amili
esb
Age
d Fa
mili
esb
All
Size
s
Fam
ily S
ize
All
Size
s
Fam
ily S
ize
All
Size
s
Fam
ily S
ize
12
34
5÷1
23
45+
12
34
51-
08.
49.
16.
98.
29.
510
.511
.513
.010
.99.
49.
910
.94.
03.
44.
34.
27.
65.
83-
512
.115
.310
.611
.711
.09.
613
.818
.413
.712
.211
.09.
66.
46.
05.
88.
111
.59.
35-
1014
.518
.215
115
.013
.611
.515
.019
.716
.415
.213
.611
.510
.38.
010
.312
.912
.711
.810
-15
15.2
18.3
17.0
16.4
14.4
12.9
15.3
20.1
17.6
16.6
14.4
12.9
12.8
8.7
12.7
14.8
15.6
13.3
15-
2015
.215
.616
.515
.815
.313
.915
.517
.617
.516
.215
.313
.912
.210
.110
.912
.817
.815
.320
-25
15.6
14.6
16.5
16.4
15.7
14.6
16.0
19.6
17.6
17.0
15.7
14.5
10.6
5.0
10.1
10.6
15.1
14.8
25-
5015
.512
.414
.316
.616
.215
.616
.017
.515
.816
.916
.015
.69.
96.
78.
313
.919
.610
.150
-10
019
.822
.217
.819
.721
.421
.021
.123
.921
.519
.921
.421
.04.
99.
64.
6d
100-
500
21.8
16.5
22.3
25.1
24.6
20.7
21.9
16.5
22.5
25.1
24.6
20.7
11.2
—10
.9—
——
500-
1,00
022
.719
.423
.223
.826
.023
.122
.719
.423
.223
.826
.023
.1—
1,00
0 an
d ov
er20
.819
.820
.323
.821
.121
.720
.819
.820
.323
.821
.121
.7—
——
——
—
All
clas
ses
15.4
16.5
16.1
16.2
15.4
14.0
16.0
18.9
17.7
16.6
15.4
14.0
9.8
6.7
9.1
12.7
15.5
13.0
a E
ffec
tive
tax
rate
s ar
e ca
lcul
ated
on
the
basi
s of
fam
ily in
com
e be
fore
tran
sfer
s, e
xclu
ding
the
amou
nt o
f co
rpor
atio
n in
com
e ta
x al
loca
ted
tofa
mili
es in
the
ME
RG
E F
ile.
bF
amili
eshe
aded
by
an in
divi
dual
age
64
or u
nder
are
con
side
red
nona
ged;
thos
e he
aded
by
an in
divi
dual
age
65
or o
ver
are
clas
sifi
ed a
s ag
ed.
CE
xclu
des
fam
ilies
with
neg
ativ
e in
com
es.
d L
ess
than
hal
f of
1 p
erce
nt.
TA
BL
E A
.5C
umul
ativ
e D
istr
ibut
ion
of B
efor
e.T
axa
and
Inco
me
for
All
Fam
ilies
, 196
6
a B
ased
on
dist
ribu
tion
of f
amily
inco
me
befo
reb
Bas
ed o
n fa
mily
inco
me
incl
udin
g tr
ansf
ers
and
(in
cum
ulat
ive
perc
enta
ges
of in
com
e)
tran
sfer
s, e
xclu
ding
the
amou
nt o
f co
rpor
atio
n in
com
e ta
x al
loca
ted
to f
amili
es in
the
ME
RG
E F
ile.
excl
udin
g co
rpor
atio
n in
com
e ta
x an
d af
ter
indi
vidu
al in
com
e an
d pa
yrol
l tax
es.
—S
Inco
me
All
Size
s1-
Pers
onFa
mily
2-Pe
rson
Fam
ily3-
Pers
onFa
mily
4-Pe
rson
Fam
ily5+
-Per
son
Fam
ily
Bef
ore-
Aft
er-
Bef
ore-
Aft
er-
Bef
ore-
Aft
er-
Bef
ore-
Aft
er-
Bef
ore.
Aft
er-
Bef
ore-
Aft
er-
Dec
ile'1
Tax
Tax
Tax
Tax
Tax
Tax
Tax
Tax
Tax
Tax
Tax
Tax
Low
est
0.31
L98
0.59
3A3
0.39
2.54
LO
S2.
671.
792.
851.
342.
63Se
cond
1.96
4.93
1.19
6.26
1.79
5.80
4.74
7.06
6.16
7.61
5.45
7.23
Thi
rd5.
7494
81.
789.
395.
0110
.37
10.0
612
.64
11.9
913
.63
11.0
113
.10
Four
th11
.49
15.5
14.
8413
.85
10.0
316
.28
16.8
619
.57
19.0
320
.75
17.9
620
.21
Fift
h18
.96
23.0
29.
8119
.86
16.9
723
.45
25.0
027
.69
57.0
928
.83
25.9
028
.22
Sixt
h28
.11
31.9
617
.31
27.6
325
.61
31.8
834
.37
36.9
236
.42
38.1
335
.12
37.4
3Se
vent
h39
.04
42.5
027
.94
37.4
736
.15
41.8
045
A0
47.3
746
.97
48.6
845
.51
47.7
6E
ight
h52
.15
55.1
341
.47
49.4
948
.74
53.4
857
8059
.73
59.2
660
.88
57.6
759
.74
Nin
th68
.36
70.6
759
.61
65.1
364
.69
68.1
272
.87
74.4
573
.93
75.3
472
.23
74.0
2H
ighe
st10
0.00
100.
0010
0.00
100.
0010
0.00
100.
0010
0.00
100.
0010
0.00
100.
0010
0.00
100.
00
Low
est 9
5 pe
rcen
t78
.70
80.5
271
.58
75.3
775
.09
77.7
282
.18
83.4
582
.96
84.1
481
.48
82.9
9L
owes
t 99
perc
ent
91.0
392
.16
85.5
687
.66
88.4
990
.19
92.8
893
.70
93.3
894
.17
92.8
393
.67
Gin
i coe
ffic
ient
of
ineq
ualit
y.4
595
.399
8.5
875
.449
6.4
958
.404
9.3
740
.332
2.3
439
.314
5.3
656
.327
9
Perc
enta
ge r
educ
tion
inar
ea o
f in
equa
lity
13.0
23.5
18.3
11.2
8.5
10.3
In c
ome
Dec
ile'2
Al!
S
Bef
ore-
Tax
b
izes A
fter
-T
axC
1-Pe
rson
Bef
ore-
Tax
Fam
ily
Aft
er-
Tax
2-Pe
rson
Bef
ore-
Tax
Fam
ily
Aft
er-
Tax
3-Pe
rson
Bef
ore-
Tax
Fam
ily
Aft
er-
Tax
4-Pe
rson
Bef
ore.
Tax
Fam
ily
Aft
er-
Tax
5+-P
erso
Bef
ore-
Tax
n Fa
mily
Aft
er-
Tax
Low
est
0.78
1.85
0.49
1.81
0.92
2.04
1.42
265
1.99
2.88
1.41
2.64
Seco
nd3.
915.
4816
64.
113.
995.
835.
477.
156.
487.
7155
97.
29T
hird
8.83
10.6
64.
587.
778.
7711
.02
10.9
612
.88
12.4
013
.80
11.1
813
.17
Four
th15
.16
17.1
49.
4213
.08
14.9
417
.39
17.8
419
.91
1943
20.9
118
.12
20.2
6Fi
fth
22.9
524
.95
16.1
519
.97
22.3
124
.84
25.9
528
.07
27.4
829
.00
26.0
428
.27
Sixt
h31
.95
33.9
524
.77
28.4
830
.95
33.4
635
.20
37.2
736
.80
38.3
135
.23
3745
Seve
nth
42.4
844
.38
35.0
838
.41
40.7
543
.09
45.7
147
.59
47.2
948
.83
45.5
847
73E
ight
h54
.79
56.5
447
.86
50.4
952
.41
54.5
058
.12
59.7
759
4760
.95
57.6
959
.69
Nin
th69
.92
71.4
463
.64
65.4
966
.71
68.4
172
.87
74.2
074
.00
75.3
372
.20
73.9
4H
ighe
st10
0.00
100.
0010
0.00
100.
0010
0.00
100.
0010
0.00
100.
0010
0.00
100.
0010
0.00
100.
00
Low
est 9
5 pe
rcen
t79
.44
80.7
473
.76
75.0
675
.81
77.2
581
.99
83.1
082
.98
84.0
581
.44
82.9
2L
owes
t 99
perc
ent
91.0
691
.99
85.2
386
.04
88.7
489
.81
92.6
193
.37
93.3
394
.11
92.8
493
.65
Gin
i coe
ffic
ient
of
ineq
ualit
y.4
099
.377
4.5
102
.457
0.4
310
.391
9.3
628
.329
7.3
385
.312
6.3
639
.327
7
Perc
enta
ge r
educ
tion
inar
ea o
f in
equa
lity
7.9
10.4
9.1
9.1
7.7
9.9
a B
ased
on
dist
ribu
tion
of f
amily
inco
me
befo
re tr
ansf
ers,
exc
ludi
ng th
e am
ount
of
corp
orat
ion
inco
me
tax
allo
cate
d to
fam
ilies
in th
e M
ER
GE
File
.B
ased
on
fam
ily in
com
e in
clud
ing
tran
sfer
s an
d ex
clud
ing
corp
orat
ion
inco
me
tax
and
afte
r in
divi
dual
inco
me
and
payr
oll t
axes
.C
Fam
ilies
head
ed b
y an
indi
vidu
al a
ge 6
4 or
und
er.
TA
BL
E A
.6C
umul
ativ
e D
istr
ibut
ions
of
Bef
ore.
Tax
a an
d A
fter
-Tax
for
Non
aged
' Fam
ilies
, 196
6
(in
cum
ulat
ive
perc
enta
ges
of in
com
e)
I
TA
BL
E A
.7C
umul
ativ
e D
istr
ibut
ions
of
and
Aft
er-T
ax I
ncom
esb
for
Age
d Fa
rnili
es,C
196
6
(in
cum
ulat
ive
perc
enta
ges
of in
com
e)
.
Inco
me
All
Size
s1-
Pers
onFa
mily
2-Pe
rson
Fam
ily3-
Pers
onFa
mily
4-Pe
rson
Fam
ily5+
-Per
son
Fam
ily
Bef
ore-
Aft
er.
Bef
ore-
Aft
er-
Bef
ore-
Aft
er-
Bef
ore.
Aft
er.
Bef
ore.
Aft
er.
Bef
ore.
Aft
er.
Decilea
Tax
Tax
Tax
Tax
Tax
Tax
Tax
Tax
Tax
Tax
Tax
Tax
Low
est
0.69
3.48
1.09
4.68
0.77
3.78
0.53
2.83
0.26
2.78
0.39
2.70
Seco
nd1.
397.
002.
189.
361.
547.
621.
536.
061.
55•6
.22
2.32
6.12
Thi
rd2.
0810
.52
3.27
14.0
52.
8911
.83
4.24
10.4
84.
3211
.08
5.86
11.3
2Fo
urth
3.73
14.7
34.
3618
.73
5.76
17.0
49.
3916
.70
8.88
16.7
210
.88
17.2
7Fi
fth
7.30
20.3
55.
4523
.42
10.2
723
.52
16.0
923
.75
14.9
923
.80
18.7
325
.53
Sixt
h13
.25
27.6
59.
4529
.69
17.1
831
.38
25.8
233
.24
23.5
932
.69
28.7
135
.41
Seve
nth
22.0
536
.98
17.0
738
.06
26.6
941
.16
38.5
544
.74
34.6
943
.34
42.1
247
.99
Eig
hth
35.5
749
.38
29.0
148
.99
40.0
553
.02
54.2
458
.93
50.3
957
.77
56.8
662
.33
Nin
th56
.77
66.4
848
.11
64.3
459
.27
68.2
572
.99
76.0
771
.03
75.8
174
.94
78.1
1H
ighe
st10
0.00
100.
0010
0.00
100.
0010
0.00
100.
0010
0.00
100.
0010
0.00
100.
0010
0.00
100.
00
Low
est 9
5 pe
rcen
t72
.60
78.9
263
.97
75.8
773
.08
79.0
784
.27
86.3
284
.39
87.0
786
.35
88,0
0L
owes
t 99
perc
ent
91.0
193
.01
87.2
191
.60
90.9
092
.89
95.9
896
.40
96.4
596
.91
96.4
896
.95
.G
ini c
oeff
icie
nt o
fin
equa
lity
.628
7.4
367
.679
9.4
097
.584
2.3
944
.459
8.3
595
.486
9.3
647
.423
1.3
309
Perc
enta
ge r
educ
tion
inar
ea o
f in
equa
lity
30.5
39.7
32.5
21.8
25.1
21.8
a B
ased
on
dist
ribu
tion
of f
amily
inco
me
befo
re tr
ansf
ers,
exc
ludi
ng th
e am
ount
of
corp
orat
ion
inco
me
tax
allo
cate
d to
fam
ilies
in th
e M
ER
GE
File
.b
Bas
ed o
n fa
mily
inco
me
incl
udin
g tr
ansf
ers
and
excl
udin
g co
rpor
atio
n in
com
e ta
x an
d af
ter
indi
vidu
al in
com
e an
d pa
yrol
l tax
es.
CF
amili
eshe
aded
by
an in
divi
dual
age
65
or o
ver.